Dolor's Rants

Gaining a competitive advantage; a value chain approach to analytics and optimisation

2013-11-28T12:04:00.000-08:00

For decades companies have been exploiting new technologies to improve their performance and to gain a competitive advantage. Over time the areas in which they want to improve and the challenges they want to address haven’t changed. Every company wants to identify its most valuable customers, its most important products/markets and wants to perform its activities most efficient. The competitiveness of a company comes from how successful it is in achieving “most valuable”, “most important” and “most efficient”. Operations Research and analytics offer new ways to achieve these goals, but how can companies determine when and how it will help them achieve their goals and gain a competitive advantage?

A company basically has two ways to improve its overall performance. First option is to improve operational effectiveness resulting in cost reductions. Operational effectiveness, most of the times, translates into standardized processes and best practices. As soon as activities become standard or best practice they are obviously going to be copied by the competition, reducing the competitive advantage to zero. Enhancing business differentiation is the second option to improve overall performance; it allows companies to ask for a premium price for the product/service, hence increasing revenue. A sustainable competitive advantage comes from doing things differently, having a different way of competing, distinctive to the company. Focus on competitiveness alone is not sufficient though; operational effectiveness is necessary in order to stay in the game.

In their drive to become “most efficient” many companies have changed their organisation to conform to ERP technology (=standardisation and best practices), instead of the other way around. As a result there is not much difference between companies in the way they make decisions and therefore there is little or no competitive advantage of using an ERP. The positive thing about ERP technology is that data on nearly every activity of a company is being recorded and available for analysis. Today, more and more companies are starting to use that data to populate dashboards (descriptive analytics) to learn about their current performance with the ambition to use predictive and prescriptive analytics to improve their competitiveness. The ever growing availability of other (external) data sources further strengthens this development. Operations Research and analytics provide the models and tools to exploit this (big) data, to find actionable insights but still have a structured decision making process. This offers the opportunity to remain efficient, but also become distinct and gain a competitive advantage.

Michael Porter’s value chain approach can help companies identify the areas in which they can most benefit from analytics and operations research. The value chain approach divides a company into technologically and economically distinct activities, each with a specific contribution to the value created by the company. The activities are spilt in two categories. Primary activities deal with the steps and processes required to transform raw materials into products or services, including sales. Next to these, activities like infrastructure and human resource management are identified to enable the primary activities.

Michael Porter's Value Chain

A company’s value chain is a system of interdependent activities, which are connected by linkages. These linkages exist when an activity affects the cost or effectiveness of one or more other activities of the company. For example, the quality of insourced oil (procurement) can have a high impact on the cost of refining it (operations). Linkages also require activities to be coordinated. On time delivery of a parcel ordered at Amazon requires procurement, inbound logistics, operations, outbound logistics and customer service activities to work smoothly together. As Michael Porter describes in his book On Competition, careful management of linkages is often a powerful source of competitive advantage because of the difficulty rivals have in perceiving them and in resolving trade-offs across organisational lines. A recent survey by PwCconfirms this as it showed that supply chain leaders, like Apple, often outsource production and delivery (both primary activities) but retain global control (they manage the linkages) over the core strategic functions such as new product development, sales and operations planning and procurement. A Gartner survey of supply chain executives in 2012 highlighted that the inability to coordinate and synchronise end-to-end supply chain processes as one of the major obstacles to gain a competitive advantage. This is exactly were operations research and analytics can offer a helping hand.

Effective decision making in supply chains, optimising linkages between a company’s activities and across companies in the supply chain, requires advanced modelling and analytical skills because of the global scale and complexity of current supply chains. Achieving the competitive advantage starts with a clear view on the current performance of the complete supply chain. Using data from the ERP system and dashboards supply chain managers now can have an instant view on the performance of any part of the supply chain. Analytics can help direct their attention to those parts of the supply chain that perform different from what is to be expected. For example, a supply chain manager at Zara can analyse point of sales (POS) data to identify and notify suppliers of potential stock-out situations and adapt the global replenishment strategy to changes in local demand. The insights drawn from this analysis directs actions to for example better manage inventory by identifying slow and fast movers, reducing stock levels and capital requirements throughout the supply chain. Using predictive models, combining actual sales figures, inventory levels, CRM data and data from social media, allows for the optimisation of sales promotions which will increase revenue as (potential) customers will be targeted more effectively with a better offer. More complex questions like selecting the number and location of warehouses are best addressed with prescriptive models, as these models are capable of evaluating and optimising all linkages in the supply chain while taking into account every relevant detail. As a result a distinctive and specific supply chain design will be identified, improving asset utilization and profitability.

We are shifting from a world in which data is arranged a head of time in ERP systems, to serve a predefined process, to a world in which data is arranged at the point of use and at the moment of use to serve the decision at hand. Smart use of optimisation models will allow companies to move away from best practises and standardized way of decision making, towards a specific decision making process opening up new opportunities to gain a competitive advantage. A value chain approach provides guidance in identifying where and how operations research and (big data) analytics can best be put to work.

The Impact of Operations Research on People, Business and Society

2013-09-13T11:52:00.000-07:00

Last week the international conference on Operations Research OR2013 took place in Rotterdam. The conference was the result of the close corporation of the GermanOR society (GOR), the Erasmus University Rotterdam and the Dutch OR society. The conference started with a keynote from Alexander Rinnooy Kan, an expert in our field and several times ranked as the most influential Dutch Person, on “How to educate Operations Research practitioners”. Key point in his lecture was that although Operations Research practitioners have been successful in applying their knowledge, continued education is of great importance. The obvious reason is keeping pace with new developments in research; on the other hand trends like Big Data give rise to new problems and applications, here research and practice can go hand in hand. The OR2013 conference programme had a special plenary session, led by the Chair of the Amsterdam Business School Marc Salomon, in which special attention was given to the impact of Operations Research on people, business and society. Over 50 C-suite business representatives attended the plenary as a special guest.

Wim Fabries of Dutch Railways

The session started with Wim Fabries, Director of Transport and board member of the passenger division of Dutch Railways. Fabries illustrated that providing reliable rail transportation is complex, requiring many interrelated decisions concerning the time table, rolling stock and crew. To support these decisions Dutch Railways has a special department, the department of Process Quality and Innovation, which uses Operations Research to support these decisions. The Operations Research practitioners of Dutch Railways had their finest hour when a new and robust time table had to be constructed which facilitated the growing passenger and freight transport on the already highly utilized railway network. The new schedule was successfully launched in December 2006, providing a daily schedule for about 5,500 daily trains. This “tour de force” was rewarded with the Franz Edelman award in 2008. Fabries indicated that Operations Research continues to be of high importance. For example in effectively managing disruptions or in constructing a reliable winter schedule so people can continue to use the trains with minimal impact on their travel plans.

Pieter Bootsma of AirFrance KLM

Pieter Bootsma, Executive Vice-President Marketing, Revenue Management and Network at Air France KLM, explained that without the use of Operations Research successfully running an airline would be impossible. In nearly every process within the company Operations Research is involved. Whether it is in strategic planning, crew management, flight operations, planning of ground services or maintenance scheduling, without Operations Research managing these processes would be nearly impossible. Given the narrow margins in the airline industry, slight improvements in efficiency can make the difference between profit and loss. In his talk, Pieter Bootsma highlighted the use of OR in Revenue management. The essence of revenue management is to use price or availability of seats to influence customer demand. By analysing booking behaviour of passengers, Air France KLM is able to estimate the willingness to pay a certain price of each passenger category. For example business people are willing to pay more for a seat than leisure passengers and they tend to book there flights closer to the actual departure date. With this knowledge Air France KLM can use the availability of seats and set the right price to maximize revenue. As a consequence availability and/or price of a passenger seat will vary over time. Revenue management has led to a paradigm shift in the airline business as it focusses on maximizing revenue, not the number of seats occupied. By many it has been coined the single most important technical development in transportation management, showing that Operations Research can be a disruptive technology.

Luke Disney of North Star Alliance

The impact of Operations Research on humanitarian assistance was illustrated by North Star Alliance Executive Director, Luke Disney. The North Star Alliance started as a practical industry response to an urgent health problem, the spread of HIV/AIDS among truck drivers in sub-Saharan Africa, negatively impacting the distribution of relief food to hungry communities. By establishing a network of drop-in health clinics, called Roadside Wellness Centres, at truck stops, ports, rail junctions and border crossings North Start Alliance can offer mobile populations like truck drivers with essential healthcare and information. The access to healthcare allows truck drivers to get treatment when necessary while at work, securing the distribution of relief food and road transportation in Africa. Luke Disney highlighted the impact of Operations Research with Polaris; a model that is used to optimise the placement of new and repositioning of existing RWCs, including the optimisation of staffing and inventory levels. Key in building the network of RWCs is to improve the continuity of care along the trade lanes within Africa. Continuity of care ensures that truck drivers can have access to healthcare everywhere they go. Also it ensures that medical help is in the neighbourhood when assistance is required suddenly, for example when a truck driver gets Malaria or when he loses his pills while being treated for Tuberculosis or HIV. Since financial resources are limited, the Polaris model helps the North Star Alliance to gradually build a network, having the biggest increase in continuity of care for each dollar invested.

Operations Research has proved to be the best answer to handle complexity many times in the past. It came into existence during the 2nd world war, where mathematicians revolutionized the way wars are waged and won by applying mathematics to almost any challenge in warfare. Today Operations Research has found its way in many applications that impacts business, people and society. The above stories point this out very clear. You could say it is the least known most influential scientific practice of our time.

Photos Anne Marie Hazenberg,

Will Big Data end Operations Research?

2013-09-01T07:41:00.000-07:00

Until a few years ago the term Big Data was only known to a limited group of experts. Now it is nearly impossible to visit a website or read an article without stumbling across Big Data. A Google query results in over 1.5 billion hits in less than half a second, two years ago this was only one fifth of that number. Hits containing links to web pages on the increased number of visitors of a museum or the improvement of the supply chain performance of Tesco due to use of Big Data. You get the impression that Big Data is everywhere. Many times Big Data is positioned as the answer to everything. It is the end of theory as former Wired editor Chris Anderson wants us to belief. The promise of Big Data seems to imply that in the sheer size of data sets there lurks some kind of magic. When the size of the data set passes some critical threshold the answer to all questions will come forward as if Apollo no longer lives in Delphi but in large data sets. Has Deep Thought become reality with Big Data and will it answer all our questions, including the ultimate question to Life, The Universe and Everything? Or a slightly simpler question whether P = NP or not? Will Big Data end Operations Research?

red = Big Data, blue = ORMS

The introduction of enterprise wide information and planning systems like ERP and the Internet has led to a vast increase in the data that is being collected and stored. IBM estimates that each day 2.5 quintillion bytes of data are generated and this rate is growing every day. So fast that 90% of the data that we have available today arose in the past 2 years. The ability to use this data can have enormous benefits. The success of companies like Google and Amazon proofs that. When sales of two items correlate at Amazon they end up in each other’s “Customers Who Bought This Item Also Bought” lists, boosting sales. The same principle is used by Google in their page rank algorithm. Using similar techniques Target was able to identify a correlation between the sales of a set of products and pregnancy. Using point of sales data and data from customer loyalty cards this correlation was used to personalise ads and offers sent to Target customers upsetting a father when his teenage daughter started to get ads for diapers and baby oil. Quite a story, but is it proof of the success of big data? We must be cautious not to be fooled by our observation bias, as we don’t know how many Target customers incorrectly received the pregnancy related ads.

As data is not objective and correlation most of the time doesn’t imply a causal link we must be cautious in blindly following what the data seems to tell us. Data is created as we gather it, it acquires meaning when we analyse it. Every analyst should know the many pitfalls in each of these steps. For example, the analysis of Twitter and Foursquare data during hurricane Sandy from the New York area showed some interesting results. The number of tweets coming from Manhattan suggested that Manhattan was worse off than Coney Island. However, the reverse was true, as smart phone ownership is much higher in Manhattan. Due to black outs, recharging smart phones became impossible, lowering the number of tweets and check-ins from Coney Island even more. A similar thing happened in the beginning of the year when Google estimated the number of flu infected people. Google estimated that 11% of the US population was infected, twice the level the CDC estimated. Cause of the overestimation probably was the media hype boosting the number of Google queries on the subject. A lot of data not a panacea after all?

Making decisions based on correlations found in data alone can bring you benefits when you are a Target customer, but can keep you from rescue when you’re living in Coney Island. Quality decision making doesn’t result from data alone, let alone the random quest for correlations in large data sets. Data however is an important ingredient for quality decision making. As Ron Howard suggests, quality decision making starts with framing the problem. The decision is supported by what you know (data), what you want (decision criteria, objectives) and what you can do (alternatives, requirements and restrictions). Collectively, these represent the decision basis, the specification of the decision. Logic (the math model) operates on the decision basis to produce the best alternative. Note that if any one of the basic elements is missing, there is no decision to be made. What to decide when there are no alternatives? How to decide between alternatives when it’s unclear how to rank them? If you do not have any data linking what you decide to what will happen, then all alternatives serve equally well. The reverse is also true, gathering data that doesn’t help to judge or rank the alternative decisions to the problem is pointless. Data is said to be the new oil. My take is that organisations shouldn’t be fixated on gathering and mining data but combine it with a structured approach on decision making. Than data will becomethe new soil for improvements, new insights and innovations.

It took Deep Though 7.5 million years to answer the ultimate question. As nobody knew what the ultimate question to Life, The Universe and Everything actually was, nobody knows what to make of the answer (42). To find the question belonging to the answer (some kind of intergalactic Jeopardy!) a new computer is constructed (not Watson) which takes 10 million years to find the answer. Unfortunately it is destroyed 5 minutes before it reaches the answer. An Operations Researcher (or Certified Analytics Professional) would probably have done a better job, starting with framing the question, gathering and validating the relevant data, constructing and calibrating a model, and finaly providing the best possible answer. I already know that it isn’t 42 or Big Data, but Operations Research.

Incroyable, fraude aux examen

2013-07-07T09:34:00.000-07:00

Copy of stolen exam

The last few weeks the papers were filled with the largest exam fraud ever in the Netherlands, the news even reached the New York Times. Three high school students were arrested under the suspicion of stealing copies of national tests of over 20 courses. The fraud was first detected at the end of May when a French Language test was posted online. To be save the board of Examination postponed the French test, affecting almost 17,000 students. It is unclear how many students benefited from the stolen exams, investigations are still in progress. As a precaution, all exams for the school at which the exams were stolen have been declared invalid and the students had to take the examinations again. But will all those who benefited from the stolen exams be found, even if they studied at other schools? Luckily, fraud can be detected using mathematical techniques and sharp thinking.

Fraud and cheating is of all times. As W.C. Fields puts it: “if it is worth having, it’s worth cheating for”. In the case of the Dutch exam fraud, since most of courses are examined using multiple choice tests, statistical test can be applied to find smoking guns for fraud. During an examination, each student fills in a scantron bubble sheet. These sheets can be read using a special device that digitizes the sheet into a string similar to:

012345678 2333324424444441241255352413511441414343113412.....

The string starts with the student id number followed by the answer string. The numbers 1,2 ... map to answers A,B ... respectively. Comparing this string with the answer key will result in the number of correct answers and the overall score. When a student wants (or needs) to cheat to pass the exam, he probably doesn’t want to change all the wrong answers, that would give a way too much certainly if his past performance is not that good. So, he needs to decide which answers to change, for example by making mistakes on purpose. It would be reasonable to expect that the more difficult the questions, the more students will get the answer wrong. Since the fraudulent student doesn’t know which questions are more difficult than others, his answers will deviate from the expected pattern as he will have more difficult questions right and more simple ones wrong. This deviation can be detected using statistical tests. In case students cooperate (or conspire), their answers will have a similar pattern. By using statistics to test for similarity of parts of the answer string these patterns can be detected as well, for example using Cohen’s Kappa statistic.

Cheating is everywhere, not only in high school exams. Using statistical tests this can be detected, even when there are no stolen exams uploaded to the Internet that act as a red flag. The tests indicate that something might be wrong and could direct further investigations. These tests don't only expose students but teachers and professors as well. In 2011 Diederik Stapel professor and dean of the School of Social and Behavioral Sciences at Tilburg University stepped down because of committing research fraud. It was his misuse of statistics that gave him away. The numbers didn’t lie, even though they were made up.

The incredible story of the expected that is not to be expected

2013-05-11T08:22:00.000-07:00

Once in a while I get an invitation from my bank to discuss my personal financial situation. Most of the time I ignore these invitations because I know their objective, it’s like this. After reviewing my investments, pension and mortgage the lovely lady (most of the time it is a lady, don’t know why) suggests to change my investment portfolio for better performance and puts a brochure in front of me of a new investment opportunity. Most prominent part of the brochure is a bar chart with the expected return of the investment compared with a couple of other investments (from other banks) and of course the new opportunity does best. But can I expect this expected return? Kahneman taught me not to trust my intuition in this, so this time I decide to go home and do some analytical experiments and decide based on the facts.

Suppose it is suggested to invest in the Dutch AEX index, an index composed of Dutch companies that are traded on the Amsterdam stock exchange. In the table below you find the annual returns over the past 30 years.

If I had invested €1,000 at the start of 1993 and reinvested the returns each year, the investment would have grown to €2,642 at the end of 2012 with an average annual return of 8.38%. The future value of the investment the bank is suggesting is calculated using the average annual return of the past and compounding it over a 20 year period. In a formula: I · (1 + r)ⁿ, with n = 20, I = the amount invested and r= the expected annual return. Following this logic and assuming that returns from the past are representative for the future a €1,000 investment now would result in €1,000 · (1 + 8.38%)²⁰ = €4,999 in 20 years. Wait a minute, how does this compare to the €2,642 we calculated earlier? The €4,999 is nearly twice as much. How come that the investments are expected to do better than the past? My experiment will show why and also that the future value of the investment is improbable because it will occur less than half of the time. So what the bank is telling me to expect is not to be expected.

A small bootstrap experiment will show the improbability of the expected value. In the bootstrap I created 10.000 sequences of 20 annual returns. Each sequence is created by randomly selecting a return from the 1993-2012 period which is placed back in the sample again, repeated 20 times. The summary results of the bootstrap are summarize in the table and graph

As you can see the expected value indeed is larger (€4,985) compared the realized value of €2,642 but also that in nearly 70% of the time the future value will be lower than average. Also note that the spread between the minimum value and the average is much smaller than between the average and the maximum value, the distribution of the future values is skewed. Cause of all this is the effect of compounding, which is key in explaining why the expected is not to be expected.

The expected value of the investment is derived by compounding the initial investment with the arithmetic average of the returns. In our case 8.38% so theoretically a future value of €4,999 results. However, the median value of the investment, the value for which there is a 50% chance that we will either fail or exceed, is derived by compounding the €1,000 using the geometric average return (or constant rate of return). The geometric average return for 1993-2012 equals 4.98% therefore the theoretical median value equals €2,642.

As you can see from the graph, the compounding of returns causes the distribution of the future values to become skewed. There are fewer values in excess of the average future value, but they exceed it with a greater amount than the amount by which the below average values fall short of the average value. So next time you are offered an investment opportunity or new pension plan be sure to realize that the expected value is an inflated estimate of the future value, the median future value is more realistic as you can expect to achieve or exceed it half the time. The expected is to be expected in that case. Besides the now standard disclaimer that past performance is not necessary indicative for future results, banks should also remark the results shown are an overestimation of the expected results.

Patterns; from tessellations to buying behaviour

2013-04-01T08:27:00.000-07:00

A few weeks ago I was invited by the director of Museum Gouda, Gerard de Kleijn, to give a lecture. Not just a lecture on optimisation but one that would link an item from the museum to my profession. There are many ways in which art and math are related with the obvious topics being the golden ration, perspective, topology and fractals. Walking through the museum with Gerard, having a look at the stained glass windows, the ceramics, the art work from the Haagse School I realised that all had to do with modelling, the kind of think I do all the time using mathematics. The closest link to math I found in the various tessellations in the museum buildings which immediately gave me the theme of my lecture, Patterns. Math after all is also known as the science of patterns. See for example Hardy in A Mathematician’s Apology.

A mathematician, like a painter or a poet, is a maker of patterns. If his patterns are more permanent than theirs, it is because they are made with ideas

Patterns are very important to us humans. Our brains are pattern recognition machines because of our associative way of learning. Thousands of years ago these pattern recognition capabilities helped us survive because our brain could recognise the sound of a predator in the grass. A peculiar thing to note is that pattern recognition is a right brain activity, the part of the brain associated with creativity and design. The left side is associated with logic and analytical thinking. Does math modelling bring these two together?

In my lecture, starting from the tessellations in the museum buildings, I explained the math that goes into designing these kinds of tessellations. It’s basic basic geometry. A few hundred years back Sebastien Truchet (1657 –1729), a Dominican Father, was the first to study tessellation mathematically. He worked on the possible patterns one could make with square tiles that were divided diagonally. His model of pattern formation was later taken up by Fournier and is now known to mathematicians and designers as Truchet tiling. Fun to note maybe is that Truchet is also the designer of the font Romain le Roi, which we now know as Times New Roman (a designer’s worst nightmare?). Dutch designer M.C. Escher, famous for his mathematical inspired designs, took tessellations much further than Truchet. He took knowledge of many famous mathematicians like George Pólya, Donald Coxeter and Roger Penrose and created wonderful woodcuts and lithographs. He became fascinated with tessellations after he had visited the Alhambra in Granada which is filled with mind blowing ancient tessellations. How those were created still is a puzzle. In Islamic tessellations not one but many types of tiles are combined, best-known are a set of tiles called the Girihtiles. Tessellations with these tiles go back to around 1200 and their arrangements found significant improvement starting with the Darb-i Imam shrine in Isfahan in Iran built in 1453. Today it is still not clear how the craftsmen at that time were able to create these wonderful and very complex tessellations as it is believed that the level of mathematical sophistication at that time was not high enough. It was only in 2007 that research showed this type of complex tessellations are comparable to Penrose tiling, predating those with five centuries.

Patterns not only appear in tessellations alone. Patterns can be found nearly everywhere. You see them in nature, in architecture, in our DNA, in fashion, in language, in music and in our buying behaviour. But instead of the clear patterns (and beauty) in tessellations, patterns in buying behaviour need to be discovered first, using advanced mathematical techniques. The basic material for pattern discovery in this case isn’t ceramic, stone, steel, or even glass but sales slips. Many grocery stores keep track of our purchases in brick and mortar stores and on line. Dutch firm Ahold for example registers over 4 billion individual items purchased each year. They have been doing so since 2000. Imagine the amount of data available to them. Until recently this data was used to develop efficient replenishment strategies for the stores and to optimise sourcing and stock levels. Using advanced mathematical techniques like pattern recognition and machine learning they are now in search of our buying patterns.

Compared to grocery stores from 60 years ago, many things have changed. In those days the shop owner knew us well and was able to offer us suggestions that with high probability would fit our preferences. But in the 1950’s shops grew larger and instead of the grocer we collected the groceries we wanted ourselves. As a consequence knowledge of our preferences eroded, leaving the grocer (and the producers) clueless on what products to offer us. Today by keeping track of the items sold and using advanced mathematical techniques the grocer is regaining that knowledge. By mining for patterns in the vast amount of data the grocer can create a picture of his clientele; which groups of customers (tribes) visit his store, what defines them, what brands they buy and what their preferences are. Based on this information the grocer can better target these tribes in marketing, either in paper, online or even in-store, also called micro targeting. (see my blog)

Mining for patterns is not restricted to analysing buying behaviour. The techniques available have become incredibly powerful in a range of fields, from the workplace to the voting booth, from health care to counter-terrorism. However, great care must be taken when using the results from the data mining algorithms. First the algorithms search for correlations in the available data. That doesn’t mean that there actually is a causal relation. A famous example is that ice cream consumption causes drowning. As ice cream sales increase, so does the number of drowning deaths. The conclusion is however wrong because it fails to recognize the importance of temperature. As temperature rises, both ice consumption and water based activities increase (and therefore drowning deaths). Second, careful attention must be paid on the data used in the analysis. It is tempting to use the data that is available, that doesn’t imply that it is sufficient. It’s like the drunk looking for his keys under the lamppost and not in the dark alley where he lost them because it’s to dark to see in the alley. Under the lamp the light is much better. So think about what needs to be analysed and gather all relevant data. These are only two of many prerequisites for successful pattern discovery. Searching for patterns really is a specialist’s task, even if your statistical package supplier wants you to think different.

Math can be used to design complex and beautiful patterns; it is also required to discover patterns in mountains of unstructured data. Math indeed is the science of patterns. Applying is can create enormous value and beauty, but in the hands of untrained users it will be useless. To be effective training is required, becoming a numerical craftsman takes time and experience like the masters whose work is displayed in the Museum Gouda.

Winds of Fortune or Despair?

2013-03-01T12:48:00.000-08:00

Have you ever been to your local grocery store and get paid at the checkout instead of paying for the contents of your shopping cart? In the European electricity markets this was the case a couple of times in the past few years. Electricity prices turned negative due to excess supply of electricity and low demand. In Europe, electricity comes from various sources. A mix of coal, gas nuclear, hydro and a rising number of wind and solar power plants provide the electricity. Current technology doesn’t allow storing electricity efficiently yet, so demand and supply must be matched at all times, with price as the principle mechanism to steer supply. On the electricity market, the lowest bidder looking to supply the grid with electricity wins. Since wind and sun come at no cost, subsidies for wind or solar power production allow for bids below €0 per unit while the electric company is still making a profit. Also because nuclear, hydro and coal fired power plants can’t be shut down without considerable cost, electric companies sometimes bid below €0 per unit, simply because the cost of shutting down and restarting is higher.

The uncertainty in electricity (spot) price is only one of the many uncertainties decision makers in the electricity market face in their decision making. The emergence of new technology increases the complexity of their decision making even more. The recent shale gas boom in the US caused electricity generators in the US to switch from coal to gas. As a consequence US coal became cheap and found its way to Europe where it displaced gas. For the future, management of electric companies in Europe are therefore faced with difficult trade-offs. The rising capacity of wind and solar power, forces them towards a flexible mix of gas and wind/solar power. The shale gas boom in the US pushes them towards the use of less flexible coal fired plants. Which portfolio is best and what should be the weights in the portfolio for each of the power sources?

Given the rising complexity of decision making and the number of uncertainties involved, electric companies will benefit strongly from adopting an optimisation based approach. In capacity investments the electric company needs to decide on number, size, and location of new power plants. Also the power source, technology and timing of the investment need to established, all under a variety of economic and technological uncertainties. It’s a complex decision, which management typically solves by simplification. Investment opportunities are intuitively evaluated on a standalone basis, considering only a couple of the most obvious uncertainties. For each investment opportunity a net present value is calculated, which is used to rank the investment opportunities. The top ones in the list are selected as the best possible opportunities and makeup the investment portfolio.

An optimisation based approach to power generation capacity investment decisions however, will allow the electric company to evaluate opportunities in a holistic and coherent manner. Instead of evaluating the investment opportunities on a case by case basis, optimisation based methods allow for connecting the forecasted cash flow of a new power plant with the forecasted cash flow of the rest of the portfolio of power plants. Using Monte Carlo, simulating the performance of the complete portfolio for future values of relevant uncertainties becomes relatively easy. Without this link it is impossible to analyse the future financial and technical performance of the electric company in a consistent manner. An optimisation based approach also allows for decision making based on facts and not biases, predispositions or beliefs because it requires an explicit definition of performance indicators for ranking and choosing between investment portfolios. By carefully analysing the portfolios that are most promising (the ones on the efficient frontier) management can find out which new power plants to invest in. If a certain investment project is part of the majority of the efficient portfolios it certainly would be a good and robust choice to invest in.

Since we can’t predict the future, certainly not for a long period ahead, the electric company must be careful not to fixate its investment strategy for the long term. Instead, optimisation based analysis allows for periodic reruns of the portfolio model and adaption of the investment strategy as conditions change over time (like increased wind power capacity, new emission regulations, shale gas boom, etc). That way the investment decision is no longer static but becomes dynamic and allows the electric companies to benefit from the winds of fortune, instead of being taken by surprise.

Back to School

2013-01-20T08:40:00.000-08:00

Each year the Dutch Operations Research society (NGB) organizes a seminar together with the Dutch Network on the Mathematics of Operations Research (LNMB). This year we organised the seminar for the 15^th time, which traditionally is held in the geographical middle of the Netherlands, Lunteren. In the past few years the central theme for the seminar has been the practical application of Operations Research in a specific area, we had themes like health care, traffic, energy, supply chain optimisation, marketing and humanitarian aid. This year we decided to take a different angle. We wanted to give our members the opportunity to upgrade their knowledge of Operations Research by offering them tutorials on new developments in OR. This idea had gradually grown from the feedback we had had on several webinars and in-company lectures on topics in OR, for example on Portfolio Optimisation (with Professor Sam Savage) and on Robust Optimisation (with Professor Dick den Hertog). These lectures seemed to fill a need to keep up with new developments in OR especially for people that had graduated from university (either on PhD or Master level) some time ago. Theme for the 15^thseminar therefore became Back to School.

Conference center "De Werelt", Lunteren

The two main subjects for the seminar we chose are Robust Optimisation (RO) and Mixed Integer Nonlinear Programing (MINLP). Each tutorial consisted of a theoretical introduction by a specialist in the field. For RO we invited Professor Dick denHertog form Tilburg University, for MINLP we invited Professor Jeff Linderoth (@jefflinderoth) from the University of Wisconsin-Madison. Their in-depth discussion aimed to explain the essence of the new developments and focus on the practical use of the techniques presented. Second part of the lecture focussed on the practical application. Marcel Hunting from AIMMS shared insights and modelling tricks in practical MINLP modelling. We closed the seminar with a lecture from Ruud Brekelmans on the use of MINLP in optimizing the Dutch dike heights, a project that has been elected finalist for the 2013 Franz Edelman Award.

Robust optimisation is a recent development in the field of optimisation, having its roots in the early 1970’s. It explicitly takes parameter uncertainty (e.g. measurement, estimation or implementation errors) into account without assuming a specify probability distribution of that uncertainty. Professor Den Hertog explained that instead of seeking to immunize the solution in some probabilistic sense to stochastic uncertainty, with RO a solution is constructed that is optimal for any realization of the uncertainty in a given set. RO starts with modelling the uncertainty region and creating the robust counterpart of the original model. Under certain conditions on the structure of the uncertainty the robust counterpart can be solved with relative ease using LP, CQP or a Conic Optimisation solver. From a practical point of view Professor Den Hertog’s advice on applying RO is to first test the robustness of the solution to the original problem with standard sensitivity analysis or simulation (a best practice in any optimisation challenge). If it is not robust use stochastic programming if the distribution of the uncertain parameter is known, if that’s not the case use RO. In applying RO one must avoid equality constraints when formulating the robust counterpart because this can cause inconsistencies between the solution to the original problem and the solution of the robust counterpart. So modelling expertise required! Professor Den Hertog concluded that Robust Optimisation provides a natural way to modelling uncertainty and that the robust counterpart makes it tractable. A wide variety of applications of RO has already been documented from which we practitioners can extract information for our own robust optimisation projects.

The second main topic of the seminar was Mixed Integer Nonlinear programming (or was it MINLP Wars?) by Professor Jeff “Obi-Wan” Linderoth. As he tweeted, it was a once in a life time experience for him. Professor Linderoth started with the relevance of MINLP, which comes from the fact that the world is not linear. Many decision problems in practice suffer from nonlinearity, think of water network design, petrochemical product blending, and oilfield planning. The need for practical solution methods to address these nonlinearities is therefore high. The normal approach to solve a MINLP is to relax the integrality constraints and construct a convex relaxation of the set of feasible solutions. Then using a branch and bound approach to solve the problem. Professor Linderoth did a lot of research to develop a new approach to MINLP, leveraging on the available MILP technology. In the lecture Professor Linderoth explained the backgrounds on algorithm engineering, (disjunctive) cutting planes and pre-processing. In general he concluded that applying “traditional” techniques from MILP in the domain of MINLP can lead to significant improvements in our ability to solve MINLP instances, a conclusion that Marcel Hunting from AIMMS reconfirmed. He showed how pre-processing and cutting planes were key in performance improvements in CPLEX. He also provided advice to OR practitioners on how to improve solver performance by applying specific reformulations in a MINLP. As an example he showed how the product of integer and real variable in a constraint can be pre-processing to get more robust solutions.

The practice of MINLP was illustrated by Ruud Brekelmans from Tilburg University. He explained the modelling and solution background of a project he has worked on to determine new safety standards for dike heights in the Netherlands. Stated very simple the decision problem was to determine the timing and sizing of the dike heightenings minimizing the total cost, consisting of the investment cost of the heightening of the dike and the economic loss in case of a flood. If you would like to know more about te project, please visit the Analytics conference in San Antonio, as it was elected finalist in the 2013 Franz Edelman award. We wish the team lots of luck, yet another Dutch Edelman award in the making?

General conclusion for the OR practitioners from the MINLP lectures was to not be afraid of MINLP. The continuous improvement in (MINLP) solvers creates new possibilities that were not available several years ago. But, even today it is necessary to come up with the right model formulation to solve your problem; you need to help the solvers where you can. Again modelling expertise required.

Given the significant rise (over 30%) in number of participants of the seminar compared to the previous years, we can conclude that the seminar filled a need. Also the feedback on the content and the speakers was great, so we can look back on a successful seminar. We decided that next year’s seminar will have the same format, topics to be announced. Suggestions for topics more than welcome. See you next year in Lunteren!

Providing a bicycle for the mind; Mathematical modelling in portfolio decision making

2012-12-27T07:59:00.000-08:00

Today's economy is driven by global uncertainties such as exchange rates, oil and energy prices, layered upon local uncertainties involving individual projects. These uncertainties create an unprecedented number of interdependent risks that hamper high quality decision making. Most Oil & Gas companies lack a consistent approach to modelling these uncertainties and risks, specifically in core decisions like exploration and production portfolio selection. Instead, they typically use averages to represent uncertainties. This leads to a class of systematic errors known as the ‘flaw of averages’.

In March 2012 Stanford University professor Sam Savage and I were interviewed on managing uncertainty in portfolio decisions by Paragon. In this interview we shared our views on portfolio optimisation and how to coherently manage uncertainties and improve decision making in portfolio decisions.

As we discuss in the interview complexity of decision making in Oil & Gas has increased in the last decade. There is a need to find new sources of oil and gas as the production level of the current portfolios is diminishing. Because these new sources of oil and gas are more difficult to reach (like deep sea, arctic or shale) new and more complex techniques are required which have high investment costs. Together with the uncertainty of finding oil and the amount of oil found, uncertainty in decision making has increased a lot. As the stakes have risen, also the risks have increased.

The Oil & Gas industry has a history of failing to deliver on promised performance. A survey of Bickel and Bratvold among petroleum engineers shows that many executives are dissatisfied with their companies’ performance, citing budget and cost overruns. They therefore require more brains per barrel to understand the impact of uncertainty and improve decision making. Prof Savage and I conclude that mathematical models will be the enabling technology that will make it easier to get a grip on the increased complexity, include uncertainty in decision making and therefore improve decision quality. Also mathematical modelling will improve consistency in decision making and reduce time to reach a decision; it will provide a bicycle for the mind in portfolio decisions.

We argue that in mathematical modelling one should focus on understanding the impact of uncertainty and how it will influence the decision that needs to be made, as it doesn’t make sense to model uncertainty if it doesn’t impact your decision. Using mathematical modelling to support decision making under uncertainty is like using a paper plane to learn how to fly. It doesn’t take very long, it’s easy and although it doesn’t look much like the real thing it teaches you the basics of flying. Instead of trying to build a model that captures all the details (engineer-like) time should be spent on building paper planes. Once the basics are understood, the level of complexity can increase and we can go for the real thing, leading to better quality portfolio decisions. Enjoy the videos.

Managing Fuel Price Uncertainty in Logistics

2012-11-04T08:48:00.000-08:00

Recently the low cost carrier Allegiant Travel announced that it would like to introduce a way of sharing the risk of fuel price fluctuations with its customers. (See Businessweek) It will offer its customers to either go for a fixed high fare or choose a lower ticket price in exchange for sharing the risk that the cost of fuel may increase before takeoff. Taking the latter option customers will pay an additional amount depending on the fuel price developments in the period between booking and flying. FAA regulations don’t allow this way of pricing at this time, but things might change. In logistics carriers typically transfer the risk of severe fuel price fluctuations to their customers, the shippers, using fuel price surcharges. The use of a surcharge on the fuel price introduces uncertainty into the shipper’s decision making and exposes it to the risk of additional costs due to rising fuel prices. This risk became apparent in 2008, when fuel prices skyrocketed, causing many shippers to exceed their transportation budgets by millions of euros. To better manage this fuel price risk, shippers need to make better decisions when selecting carriers.

In a typical supply chain one or more carriers are used by a shipper to transport materials from manufacturing centers to warehouses, distribution centers and/or retail outlets. In selecting a carrier a shipper invites carriers to quote prices at which they are willing to haul loads on the lanes in the shipper’s network. Based on the quotes the shipper selects the carriers and signs a contract with them for a certain amount of time. This contract contains the agreement on the price per shipment the carrier asks for transporting the loads, consisting of a base price plus a fuel surcharge. In most cases a shipper estimates the increase in cost due to the fuel surcharge using the average fuel price of the previous year maybe with a small mark up. This is why things can go very wrong.

Let me illustrate with a small example of a shipping company that wants to organize its logistics between Amsterdam, Berlin, Paris, Madrid and Rome. Each week it is required to transport material between each pair of cities. The shipper has asked 3 carriers for quotes for the lanes in scope. The cost per kilometer, including the surcharge for fuel price is shown in the graph. For small distances (like Amsterdam-Berlin) a higher cost is incurred, than for long distance hauls. The shipper selects its carriers based on the average fuel cost of 2009 (€0.89/Liter) with a markup of 15% (resulting in €1.023/Liter) to capture the uncertainty on the fuel price. Bases on this business rule the Shipper will select Carrier 1 for its long hauls (>1000 km) and Carrier 3 for its short hauls. Now take a look at how the logistic cost develops during 2010.

Although the shipper used a markup on the fuel price to try and capture the fuel price uncertainty, it still faces an increase of weekly shipping cost during 2010 of over 4%. This is due to the way the increase in fuel price kicks in on the line haul cost charged by each of the selected carriers. At a fuel price of €1.08 and above, Carrier 1 even becomes cheaper on the short hauls than Carrier 3 which was selected initially. But at that time the shipper can no longer switch carriers. With hindsight it would have been better to select Carrier 1 for all line hauls.

Fuel prices are uncertain and cannot be influenced by shippers. But they can improve the quality of their decisions by better incorporating fuel price uncertainty. A straight forward way would be to test the robustness of their choice of carriers under various fuel price assumptions, or use more advance modeling methods like Monte Carlo simulation or Robust Optimization that explicitly incorporate the fuel price uncertainty when identifying the best choice of carriers. Also, knowing the negative effect fuel price can have on shipping cost, the shipper can negotiate better conditions on how fuel surcharges are imposed. An example could be to put a cap on the fuel surcharge at a certain fuel price level, so further fuel price are at the carrier’s risk. Again mathematical models can help both the shipper and the carrier decide what would be a fair way of sharing the risk. This of course depends on their risk appetite and the possibility to diversify this risk.

Would you entrust your daughter to a decision professional?

2012-10-19T09:10:00.000-07:00

Last week the European Decision professionals (EDPN) had its first conference. Theme of the conference was competing through quality of strategic decisions. Reidar Bratvold, one of the speakers, touched upon the fact that we as humans are unfit to deal with uncertainty in reasoning. The examples he used, reminded me of a book that I recently read, The Drunkard's Walk: How Randomness Rules Our Lives by Mlodinow. Bratvold had asked decision professionals in Oil and Gas industry about a well-known puzzle in statistics which is also in Mlodinow’s book. It goes like this:

You are told that a family, completely unknown to you, has two children and that one of the children is a daughter. What is the chance that the other child is also a daughter?

About 80% of the people Bratvold asked responded that the chance is equal to ½. Bratvold repeats the question making explicit that no statement whatsoever is made on the birth order of the children. With that remark, still 80% of the people responded that the chance is equal to ½.

The correct answer however is 1/3. With two children in a family there are obviously four possibilities: {Girl, Girl}, {Girl, Boy}, {Boy, Girl} and {Boy, Boy}. Since one of the children is a girl, the {Boy, Boy} possibility must be eliminated. That leaves us with 3 possible options. The chance that the other child is a girl as well therefore is 1/3. To come to the correct answer I used Cardano’s method. As he explains in “Book on Games of Chance” it is best to construct a sample space to calculate the odds of even the simplest events. Trust your instincts instead and you’re bound to go wrong, even the decision professional.

To test your instincts, what if we add a seemingly irrelevant remark to the above question? Suppose that the daughter is born on a Friday. What are the chances of both children being girls?

Let’s follow Cardona’s advice and built a sample space. We had three possible combinations, {Girl, Girl}, {Girl, Boy}, and {Boy, Girl}. For each combination count the possible combinations of weekdays of birth:

{Girl, Girl} = {Mon, Fri}, {Tue, Fri}, {Wed, Fri}, {Thu, Fri}, {Sat, Fri}, {Sun, Fri}, {Fri, Mon}, {Fri, Tue}, {Fri, Wed}, {Fri, Thu}, {Fri, Sat} {Fri, Sun}
{Boy, Girl} = {Mon, Fri}, {Tue, Fri}, {Wed, Fri}, {Thu, Fri}, {Fri, Fri}, {Sat, Fri}, {Sun, Fri}
{Girl, Boy} = {Fri, Mon}, {Fri, Tue}, {Fri, Wed}, {Fri, Thu}, {Fri, Fri}, {Fri, Sat}, {Fri, Sun}

The {Fri, Fri} outcome for the combination of two girls is not valid, because we know that only one girl was born on a Friday. In total there are 26 possible outcomes of which 12 are favorable. So the chance of two girls in this case becomes 6/13.

My and probably your intuition as well would have been to discard the additional information on the weekday of birth as irrelevant. But it isn’t! The probability even rises because of it, with whopping 38%! Imagine the impact of this is when it comes to real world decisions like exploring new oil fields, medical decisions or investing in the development of new drugs. So my advice would be before you entrust your daughter to a decision professional to ask him if is knows Gerolamo Cardano. As irrelevant that question may be, it can have a large impact.

Fair Elections?

2012-08-31T13:36:00.000-07:00

Elections for Dutch parliament are only a couple of weeks away. The papers, television, internet and social media are swamped with election rhetoric, opinions and polls. Not much different from the election circus in the States I guess. For the Dutch this election is of high importance. The national economy is under pressure because of rising national debt, sluggish domestic demand and the housing market slump. The next government needs to come up with measures to control and reduce the national debt without hurting economic growth. Which government that will be is up to the voters. But will an election result in a government that represents the preferences of voters? There are many examples of voting systems in which that is not the case, and I am not talking about elections in a banana republic.

To give an example; in 2005 the Labour party got 57% of the seats in the British House of Commons with only 36% of the votes. Is that fair? This outcome is a consequence of the first-past-the-post electoral system, or winner-takes-all system, in multiple districts. In this case, the candidate that gets the most votes, whether he/she reaches a majority of votes or not ("first past the post") wins the seat in parliament. It was designed to support for a two-party electoral contest. It worked well when there were only Whigs and Tories it is however ill-suited to a multi-party political landscape. It thwarts the will of the voters, leaving millions without political representation in parliament.

In the Netherlands the House of Commons is elected using a system of open party lists and only one district (the whole nation), resulting in proportional representation. Question is if this results in a fair representation of the voters preferences. Unfortunately this is also not the case. In 1994 the majority of the voters preferred the Democrats 66 party (D66) to the Dutch Labour Party (PvdA), Christian Democratic Appeal (CDA) and People's Party for Freedom and Democracy (VVD) so you would expect that the D66 would also have the majority of the seats in the House of Commons. However PvdA, CDA, VVD all ended up with more seats that D66. Is this fair?

Take a look at the vote and seat distribution of the 1994 elections. As you can see, CDA and PvdA suffered great losses, D66 and VVD won a lot of votes. Now consider the election matrix as retrieved prior to the elections (source Dutch Parliament Elections Studies). The cell (D66, PvdA) in this matrix contains the number of 655 while the cell (PvdA, D66) contains the number of 580. This means that 655 respondents strictly prefer D66 to PvdA and that 580 strictly prefer the PvdA.

Using this matrix a majority ranking can be constructed in such a way that the party higher in rank holds a majority over all parties at lower ranks. From the majority ranking it is immediately clear that although D66 has the majority vote it is only ranked fourth based on number of seats. This is an example of what is called The More-Preferred-Less-Seats Paradox.

Can’t we improve voting systems to overcome these kinds of paradoxes? Unfortunately not. Like Kurt Gödel's proof that there will always be facts which cannot be proved or disproved in any mathematical system, the Impossibility Theorem on social choice of Arrow precludes the ideal of a perfect democracy. So abolish democracy? Of course not! The perfect democracy doesn’t exist, the challenge is to find the best possible. With Arrow’s theorem available a trade-off can be made on the properties a voting system should have and whether the above paradox is acceptable or not. But how to decide on that? Take a vote?

May the best team win? Chances are they will not

2012-06-24T05:38:00.000-07:00

In an attempt to understand and come to terms with the early exit of the Dutch football team from the European Championships I started to review the team’s performance during the qualification phase. They must have been a poor team in that phase as well. But I found an amazing track record. The Dutch team scored 37 goals in ten matches, which was the highest of the 53 teams in the competition. Ranking all the teams of the qualification phase the Dutch came in second with 27 points, just after Germany who totalled 30 points. During the qualification phase we only lost one game, at that time we already had qualified for the finals. How come such a great team, with top players like Wesley “The Sniper” Sneijder and Klaas Jan “The Hunter” Huntelaar, didn’t make it to the finals? They didn’t even manage to score a single point in the group phase. I found the answer in basic probability analysis.

The European Football Championship starts with a pool phase of 4 groups. In each group 4 countries play against each other to determine first and second in the group. After the group phase the knock out phase starts with quarter finals, semi-finals and finals for the 8 remaining teams. The Netherlands was up against Denmark, Germany and Portugal. Given the tough competition, this group was called the group of death. Trying to figure out what the probability of surviving the group phase was I started with some basic calculations. To estimate the probability of winning, losing or a draw when playing against these countries I looked at their ranking in the UEFA list . Germany and the Netherlands score about the same (40860 for the Netherlands, 40446 for Germany) Based on this score I thought it reasonable that the probability of a draw when playing against Germany would be about 50%. I set the probability of winning or losing from Germany at 25% each. Denmark and Portugal both had a much lower score than Germany and the Netherlands in the UEFA list. A reasonable probability of wining from the Danish and Portuguese team would be 40%, losing the game at 30% with the same probability for a draw. Note that the probability of winning all the games in the pool phase for the Dutch team than becomes 40%*40%*25%, or only 4%. Based on these assumptions the probability the Dutch would survive the pool phase (for simplicity reasons, scoring at least 6 points) is 28%. Thus surviving the pool phase for the best team (The Netherlands has the highest UEFA ranking of all countries in the group) is slightly better than one in four; this is similar to drawing a random name from a set of four. It’s a lottery.

Assuming the Dutch team would have managed to get to the knock out phase, the teams they would have been up against would be the strongest in the competition. The probability of winning a game would have been similar to the flip of a coin. Thus the probability of making it to the finals would have been 28%*50%*50%=7%, winning the championship even worse, only 3.5%. Thus, although the Dutch were one of the favourites, qualifying as second for the final round of the European Championship, it was tough luck that led to their early exit. As basic probability analysis shows, the probability of the best team not winning the championship is high; winning the championship is more like wining a lottery. This insight helps to sooth the pain a little and not only supports me, but any country that didn’t make it to the finals.

What does it take to win a Super Bowl?

2012-04-25T11:48:00.000-07:00

The scene: April 16^th, Regency Hyatt Grand Ballroom in Huntington Beach, California. INFORMS President Terry Harrison walks towards the podium on the stage and says: “And now the moment of truth. Srinivas, may I have the judges’ decision please? Ladies and gentlemen, please welcome to the stage to receive the first place 2012 Franz Edelman Award for outstanding achievement in Operations Research, the team representing……..TNT Express". At table 23 there is excitement and cheers. We did it! We won the Franz Edelman Award 2012!

Winning the Franz Edelman award really is amazing. I know the work of previous Edelman Award winners like Sloan-Kettering, Hewlett Packard, the Netherlands Railways and Midwest ISO. Being ranked among these top OR accomplishments is a great honour to me. The Edelman rewards outstanding examples of applied Operations Research. It’s the Super Bowl of OR, and I’m very proud that this year the team of TNT Express was selected first place. It’s not what we set off to accomplish when we started our first project at TNT Express back in 2005. That first project didn’t even contain much OR.

The first time we considered applying for the Franz Edelman Award was at the beginning of 2011. Several of Hein Fleuren’s colleagues at Tilburg University and some of my colleagues at ORTEC had suggested that we should submit our work with TNT Express for the Edelman Award competition. At first we had our doubts. Did the OR we had used at TNT Express match the level of sophistication of past Edelman Award winners? Were the accomplished savings significant enough? However, what appealed to us is that the Edelman Award rewards implemented work. After seven years of hard work on different levels of decision making within TNT Express a lot had been accomplished. After discussing it with TNT Express and thinking about it, we decided to go for it.

Applying for the Edelman Award gave us the opportunity to share TNT’s approach in applying OR and why it was successful. Of course it requires sound OR. As an OR consultant you have to construct the right model at the right level of detail to support the business decision at hand, an art in itself. However, that’s not the complete story. A key aspect of TNT’s success in my opinion is that from the start TNT had a strong focus on the organisational embedment of the programme, a reason for many projects (not only OR projects) to fail when that’s not taken care of. This embedment was accomplished by not just focussing on tools and technology (the OR) but also with strong project- and change management principles in place and a focus on the development of analytical skills and supply chain thinking. Marco Hendriks, Hein Fleuren and Jan Salomons developed a special training programme together with Tias Nimbas Business School to achieve that, the GO-Academy.

When we started the Global Optimisation (GO) programme in 2005, TNT Express wasn’t accustomed to using mathematical models or formal methods to support decision making. Based on our experience with other customers we started with easy to understand analysis tools that supported the visual identification of bottlenecks and provided suggestions to resolve them. As the awareness and understanding of optimisation methods increased at TNT Express, we also increased the level of sophistication of the models, making sure to keep pace with the decision making maturity level of the TNT operating units. That way, decision making quality improved and OR models and tools became standard equipment for decision making. One of our biggest challenges was in 2008 when TNT Express faced major challenges as a result of the financial crisis. Shipping volumes had gone down drastically which required immediate action. Being aware of the GO accomplishments thus far, Marie-Christine Lombard asked Hein Fleuren to work on this challenge. Supported by OR modelling and analysis, TNT Express was able to counter the challenges of 2008 and reduced costs significantly with minimal impact in service. With this success, OR modelling became a key asset for senior management decision making. Today, fact-based decision making capabilities have become available at the core of TNT’s business, the primary objective of GO.

In preparing for the Edelman competition, we worked as a team, as we did in the GO projects. Winning the award strengthens our bonds even more. Standing on the stage, receiving the cheers and applause really does something to you. Not to mention all the emails, text messages and tweets with congratulations. It proves that the GO programme really is special. Reading the takeaways from the conference on the internet, one appealed to me very much. Rob Ende states: “My biggest takeaway from their (TNT Express) presentation had nothing to do with their network optimization models. Rather, it was how they built an entire optimization “ecosystem” centred on their GO (Global Optimisation) Academy.” I guess that is what it takes to win a Super bowl, you have to live what you want to accomplish. With GO-Academy and Communities of Practice, we were able to build an ecosystem that allowed optimisation to flourish within TNT.

On Eggs and Baskets

2012-03-11T08:24:00.000-07:00

Reading the papers the last couple of weeks, the reported rise of investments in both the chemical and the oil & gas industry caught my attention. The top 20 European chemical companies had doubled their investments from 1.5% of total revenue in 2008 to more than 3.1% in 2011, which is about €5 billon. The oil & gas industry also displays an increase of investments driven by the race for new production wells. Petrobras is investing a stunning $225 billion in exploration of new oil fields over the period 2011-2015, making it one of the world’s biggest corporate investment programmes. Not only Petrobras, other major oil companies report a significant rise in investments as well. Royal Dutch Shell for example reported a€12 billion investment to explore the Prelude-gas field in Australia. Chemical and oil & gas companies have to make numerous investment decisions, each of them concerning large initial investments and uncertain returns. Business executives in both industries therefore require rigor in decision making when deciding on their capital investments. One lesson they all know too well is to make sure that they diversify their investments and not place all of their eggs in one basket. But how do they decide which baskets to invest in and how many eggs to put in each of them?

Diversification of investments is not an idea that came out of the modern portfolio theory as Harry Markowitz developed it. The knowledge was already available 3000 year ago as the book of Ecclesiastes (about 935 B.C) advices you to "divide your investments among many places, for you do not know what risks might lie ahead". It’s even part of classical English literature. Antonio in Shakespeare’s Merchant of Venice tells Salarino that his ventures are “not in one bottom trusted, nor to one place”. The work of Harry Markowitz however gave us a formal method for deciding on the best possible portfolio of investments, minimizing risk given a required return. He was rewarded the Nobel Prize for Economics in 1990 for this work. Key in the work of Markowitz work is the concept of diversification which reduces the risk of an investment portfolio, but can appear counterintuitive.

To illustrate imagine two projects, a safe and a risky one, with independent probabilities of success. Each project requires an initial investment of €10 million. The table above summarizes the probabilities and pay offs for each of the projects. Note that the expected pay off for each of the projects is €30 million. Investing in the higher risk project will not increase the expected return, so the safe project is the obvious better choice. Now suppose that it would be possible to split the €10 million investment in two and invest €5 million in the safe project and €5 million in the risky project. Would that be a better choice? Intuitively it would seem like a bad idea to take money invested in the save project and invest in the risky project and I expect that many business executives also follow that instinct. However, diversification of the investment, spreading the eggs over the two baskets, will reduce overall risk. Let me show you how. Instead of just two outcomes, we now have 4 possible outcomes as summarized in the table below. The expected value of the 50/50 split remains €30 million but compared to investing solely in the safe project the probability of losing €10 million is reduced from 33% to 22%, a significant reduction of risk. Although moving money from a safe investment to a more risky one seems counterintuitive, doing so will reduce risk showing the effect of diversification.

Note that when the projects would have been positively correlated (for example when drilling for oil in the same area) this risk reduction would have been less significant. If the safe project fails, because of the positive correlation, the risky project has a greater probability of failure resulting in a higher than 22% probability of losing €10 million. On the other hand if they would have been negatively correlated the risk reduction would have been even more significant. In deciding on which projects to invest in, the executives in the chemical and oil & gas industry therefore need to spread their investments seeking the negatively correlated projects, and avoiding the positively correlated ones. A challenge that can be solved well with the optimisation techniques of Operations Research.

A Billion in Need

2012-02-05T11:31:00.000-08:00

A red cup makes the difference between life and starvation for one billion people every day. One of every seven people on earth suffers from chronic hunger, every 10 seconds a child dies of hunger. These are horrifying facts, which become even worse when we realize that there is enough food and technology available to feed everybody. Last November I was in Rome, at the head office of the World Food Program (WFP), the world’s largest humanitarian agency fighting hunger worldwide. My visit coincided with the 50^th anniversary celebration of WFP. I listened to Josette Sheeran, Executive Director of the World Food Program (WFP), and several other officials talk about the work and the challenges of WFP. At that meeting my already strong belief (see my earlier posts) that Operations Research can contribute to fighting global hunger became even stronger.

WFP manages a global supply chain for food aid, moving food to where it is needed most. This can be emergency response due to manmade or natural disasters or instances of chronic hunger. Every day WFP fills the red cups of over 90 million people in more than 70 countries. It is a complex task, in which decisions need to be made fast and with confidence since lives are at stake. The complexity of the WFP operations is, as Van Wassenhove describes in Humanitarian Logistics, comparable to planning an event like the Olympics. But imagine planning the event not knowing when or where it will take place, how many spectators will attend or how many athletes will compete. The near impossibility of this task gives some insight into what the WFP is up against.

During my visit, I talked with several WFP officials and learned in more detail the challenges they face in managing their supply chains. WFP recognizes that logistics is the part in their supply chain that can mean the difference between a successful or failed operation. It also is the most expensive part; in emergency relief 80% of the cost is due to logistics. WFP must design and manage its logistics in such a way that they get the right goods to the right place and distribute to the right people at the right time. These decisions are more complex than in private sector logistics, because WFP operates in a highly uncertain environment and has many stakeholders (Donors, Military, NGO’s, local government, etc). Things become even more complex; road infrastructure can be a serious problem. In South Sudan (a country as large as France) no more than 4,100 km of all-weather roads are available. For comparison, neighboring Kenya has 160.000 km’s of roads. The absence of all-weather roads can be a serious bottleneck when transporting food into a country, especially when the country is land-locked and no other modalities are available to get the food in.

To feed a billion people WFP needs to be successful at designing and managing its supply chain. It has to act fast, make decisions in a complex and uncertain environment with scarce resources. It is an environment in which Operations Research is at its best. With our analytical skills we Operations Researchers can develop and apply tools and techniques to solve the decision problems WFP faces with the rigor they require. From past projects I already know that this is true. Following my visit to the WFP head office we started a project to support the South Sudan operation of WFP. Our aim is to increase the decision power of WFP in the supply chain design for South Sudan, working towards a generic model that will support decision making in organizing aid in land-locked countries. With support from the Operations Research society WFP can become better, smarter and faster in providing aid to those who need it most. Let’s get to work and make sure that there is no need to celebrate the 60^th anniversary.

Losing weight fact based

2012-01-04T09:06:00.000-08:00

Made any New Year resolutions this year? What’s your #1 on the list? I bet it is losing weight. In the Netherlands it is the number one resolution for 2012 and I expect in many other countries as well. Research from ING Banking & Insurance indicates that about 80% of the Dutch have made New Year resolutions this year. By equating the fulfillment of a resolution to an economic value ING was even able to calculate that on average the Dutch would give €450 to keep their resolution, resulting in a total economic value of €4.5 billion in the Netherlands alone. Given the economic crisis we are in, a serious amount of money. To stimulate keeping the resolution, maybe the government could introduce a tax in case of failure. This could be a very interesting idea, since not many of us seem to be able to keep our resolutions. To illustrate, 88% will fail to stop smoking, 95% will keep the same weight or even gain weight.

So you want to lose weight, but how much? One way of making that estimate is to use the Body Mass Index. It is a much used number in the medical profession to measure if you’re overweight. It divides your weight in kilograms by the square of your length in meters. So if you’re 1.86 meter tall and weigh 92 kg, a BMI results of 26.6. A BMI score lower than 18.5 kg/m2 implies underweight, above 25 kg/m2 overweight. Interesting to note is that the BMI would put Schwarzenegger (in his Conan years, 1.83 m tall and about 107 kg) in the serious overweight category. I can’t imagine that to be correct. Let’s take an analyst view at this way of assessing overweight.

To calculate the BMI, only height and weight are required. Would that be enough to decide if a person is overweight? It’s important to know that the density of fat is less than that of muscle which in turn has a smaller density than bone. In other words, the less fat you have, and the more your body is made up of muscle and bone (meaning you’re a fit person), the greater the numerator in the BMI formula, and therefore the higher the BMI. So when you start to exercise to lose weight and build up muscle, you will be worse of according to the BMI. This can’t be right, the BMI model must be wrong.

The concept of BMI is the work of a Belgian mathematician, Adolphe Quetelet . He was one of the first to use statistics to draw conclusions about society. He published his "Quetelet Index" in 1832, later known as BMI. Quetelet had no interest in studying overweight when he developed his index. His main interest was to apply probability calculus to human physical characteristics which led him to develop the BMI formula. He found that during normal growth, weight tends to increase in relation to height in meters squared. So for a population as a whole, the BMI is an easy to calculate number that helps study weight issues. There is no rationale or medical evidence that explains why the model is right on indicating whether you are overweight or not. So the BMI was initially developed to measure a societally trend, not as a diagnostic tool to draw conclusions on an individual level. But that is the way in which it is now used, even by doctors. Statistical speaking, that’s rubbish.

So if not the BMI, than what indicator should you use to determine the weight you should lose? You could take a look in the mirror, or ask your partner. But for sure these are not very objective measures. So maybe the best measure is to put on your favorite jeans and feel if they still fit, not to tight. That probably is the best model to use to optimize your weight. Keeping your resolution will also become easier, because buying a complete new wardrobe will probably cost you more than €450.

Optimizing the Human Resource Supply Chain

2011-12-30T11:34:00.000-08:00

Reading the various outlooks for the coming year, I came across the Manpower Employment Outlook for 2012. Interesting statement from the report is that although employers are more inclined to hire new personnel there continues to be uncertainty about the market, so they are reluctant to make the investment in a permanent hire. At the same time employers also have trouble filling vacancies for specialist jobs like technicians, sales people, skilled trade workers and engineers, as indicated by another report from the same agency. This makes balancing the demand and supply of human resources a though job. Overstaffed units will put pressure on the EBIT because of too high cost levels, while understaffed units will suffer from service degradation or loss of revenue. However, when companies link their strategic workforce planning with their business planning the workforce productivity can be optimised which will result in a decrease the total cost of human capital, while maintaining or increasing the overall quality of their workforce and customer service.

Effectively managing demand and supply for human resources requires a structured planning process, starting with a forecast of the demand for human resources as derived from the business planning. Next the supply of resources needs to be forecasted. Since people change jobs, retire, get hired or fired available human capital will change over time, dynamics that need to be taken into account in strategic workforce planning. Matching the forecasted demand and supply for human resources indicates where shortages of overages can be expected. Making a choice on the instruments for adjusting human resource supply a company can than take appropriate action to establish the best possible balance between supply and demand. In this process the HR manager can benefit a lot from the Operations Research models in making these decisions fact based.

When comparing the dynamics of human resource capacity in supply chains with other resources in the supply chain you will see that human resources are different. The dynamics of other resources are mostly restricted to ordering, whereas human resources have a wide variety of characteristics which all influence the availability of human resources over time. To name a few; acquiring new skills, productivity increases due to learning, change of role/function and getting hired or fired. These characteristics will influence the available resource capacity and cause forecasting resource availability to be difficult. Operations Research offers all kinds of methods to incorporate these dynamics, improving the quality of forecasted availability. For example, stochastic loss network models or somewhat simpler the Markov approach as described in my The OR in HR blog entry. The parameters of these models, for example transition probabilities, need to be estimated based on the data in the HR systems and can best used together with subject matter experts to incorporate factors that are not present in the available data. These models than can be used stand alone or for optimisation purposes, like in deciding on the most cost effective capacity deployment.

In long term capacity planning it is decided how to deploy the expected available capacity, which is not straightforward to accomplish. Human resources differ from ‘normal’ resources since they are not consumed in the make process (they deliver a service) and productivity and efficiency depends upon their workload and utilization (Parkinson’s Law and the Student Syndrome in action?). Also human resources can perform more than one skill at a time and doing so across multiple assignments. In the service industry, like in Operations Research consulting and IT services, the process of assigning people to tasks/roles is further complicated by simultaneous allocation of multiple resources and resource sharing over assignments. I know from own experience that making a long term capacity plan for my team is a hard. Assignments simply don’t start when you would want them to start leading to over-utilized or underutilized consultants. Also you want assignments to fit with the capabilities of the consultant and stimulate the development of new capabilities and knowledge. The right assignment for that is not always available. To complicate things even more, both demand and supply for human resources are uncertain. In long-term capacity planning the exact assignment of resources to projects or tasks is not required. What is needed is to verify if there is enough capacity to satisfy demand or master plan. One way to answer that question is by making a rough cut capacity plan, RCCP. Usually this requires formulating and solving a mixed integer linear program. The RCCP will indicate if the master plan can be satisfied (is it do-able?) and what are possible bottlenecks and mitigation actions when the master plan changes, for example due to changes in demand.

Operations Research will improve managing the balance between supply and demand for human resources significantly. It allows for the incorporation of the dynamics of human resources in the decision making, resulting in better quality and fact based decisions. It will support management in making a choice on the instruments for adjusting human resource supply on the longer term optimizing the human resource supply chain.

The incredible balancing act of Unsold and OutOfStock

2011-11-06T07:56:00.000-08:00

Inventory management is one of the key factors determining the performance of a supply chain. A small change of the inventory policy can lead to a dramatic alteration of the supply chain’s efficiency and responsiveness. Traditionally inventory management is challenging because it directly impacts both cost and service. Uncertain demand and uncertain supply make it necessary to hold inventory at certain places in the supply chain to provide adequate service to customers. As a consequence, increasing inventories will increase customer service and revenue, but also increases cost. According to the 22^nd annual state of logistics report (pdf), the world is sitting on roughly $8 trillion worth of goods held for sale, and nearly $2 trillion in the U.S. alone. That's a lot of capital tied up in warehouses. Besides being a huge capital absorber, inventory also represents a tremendous amount of environmental footprint. If we could permanently reduce the amount of product sitting idle, we'd save money, energy, and material.

Effective inventory management is very hard. A recent article in the Financial Times illustrates this. Following the earthquake in Japan, many tech companies started to stockpile critical components to avoid shortages later on. However, unexpected low customer demand in US and Europe resulted in tremendous inventory levels. The components for which it was expected that there was a risk of shortage are now having the biggest problems in terms of oversupply. Deciding on the right level of inventory therefore is an incredible balancing act of unsold and out of stock where Operations Research can be your balancing pole supporting you in making the trade-off.

To illustrate, the above figure shows three warehouses and four customers. The supply chain manager has to decide how much stock to keep at each of the warehouses and which customer to serve from which warehouse. Stock keeping cost is given for each warehouse. Transportation cost is presented in the table below.

The supply chain manager is uncertain about customer demand but has to decide immediately on the number of stock keeping units at each warehouse, otherwise the available floor space will be leased to another company. He decides to go for the average demand. Using his MBA skills he optimises for minimal supply chain cost (warehouse and shipping cost) and decides to have 6215 SKU at warehouse 1 and 3000 SKU’s at warehouse 3. Perfect, ….or could he have done a better job? Actual demand will deviate from the average demand (average doesn’t exist, does it?) leaving the supply chain manager with either unsold stock like in the examples above or lost sales (on average 431 of either unsold or lost sales given the demand scenarios). Would minimizing for lost sales be a better option? Total supply chain cost will rise for sure, possible also increasing the level of unsold stock. Here Operations Research can be of assistance to find a balance. Using the available demand scenario’s (min and max in this case) the average total supply chain cost can be determined while varying the level of average lost sales (see graph below). Note that the supply chain cost of the average demand is not the same as the average supply chain cost of the demand (the flaw of averages!).

This way the supply chain manager can make the trade off between supply chain cost and lost sales leading to better quality decisions. He/She can use this information to share with all players in the supply chain, like marketing and sales, production and procurement, building a shared view and plan. Operations Research will help find the balance between Unsold and OutofStock and keep it when incorporating deviations from forecasted supply and demand, enabling you to practice Sales and Operations Planning fact based.

Deciding on Lean or Green

2011-10-04T12:29:00.000-07:00

Imagine getting into your car, entering your destination into the satellite navigation system and getting not only the two obvious options for the shortest or fastest route, but also the most sustainable one. What would you think of that? Calculating the shortest or fastest route is easy from an Operations Research perspective, just use Dijkstra’s shortest path algorithm. How about the most sustainable one? Whether a route is sustainable or not depends on many factors; maximising sustainability therefore is different from minimising travel time or distance. There is however a linear relationship between fuel consumption and CO2 emissions. So when fuel consumption can be taken into account, optimising for the most sustainable route would become possible. That is exactly what two of my colleagues at ORTEC, Goos Kant and Patrick Schittekat, did when researching the net effect of focussing on sustainability in logistics.

Fuel consumption is influenced by factors like engine type and aerodynamics of your car, the type of fuel used, traffic density, variance in driving speed, the weather and not to mention your driving habits. Some of these factors can be modelled easily while others, like the weather, are more challenging. Research as reported by UK’s National Atmospheric Emissions Inventory (NAEI) shows that CO2 emission levels can be expressed as a function of vehicle type, fuel type and engine type and of course speed. So when we know these parameters, we are able to calculate emission levels.

In calculating an optimal route, a digital road network is required. Companies like NavTeq and AND provide these maps. A digital road network consists of points and segments connecting these points, representing the road network in the real world. The segments are of different types, representing different road types each with a different speed. Think op motorways, regional roads, local roads and city areas. Given the length of each segment and speed of a vehicle on each of the segment types, the shortest and fastest route can be calculated. An interesting new development is that more and more information is added to digital networks, like the time/day dependent average speed at which these road segments have been travelled, allowing for a better estimation of actual travel time and therefore speed.

Using the formula for CO2 emission levels, the sustainability cost for each segment in the road network can be calculated, using the travel times reported in the digital network segments. Using Dijkstra’s algorithm, the most sustainable route (lowest emission levels) between each point in the network can be calculated and a comparison can be made with the shortest and the fastest routes. Research of my colleagues Goos and Patrick indicate that the greenest route is about 5% slower than the fastest, while the shortest is 35% slower. Also the greenest route is 2% longer than the shortest, while the fastest route is about 6% longer. Comparing on costs (using social cost per ton CO2 as reported by DEFRA and associated cost for vehicle and driver) the greenest route is about 1% more expensive than the fastest, while the shortest is 17% more expensive. So in short taking the green road home is only slightly slower and costs a little bit more than the fastest (time is money after all).

In logistics making the trade off’s between lean or green isn’t common yet but will change in the near future. A example of this trend is the special programme (Sustainable Logistics) in which Dutch companies have committed themselves to achieve a 20% reduction in CO2 emision levels by the end of 2012. Incorporating emission levels in logistic optimisation models will help create insight and guide companies towards more sustainable choices. By the way, focussing on efficiency (cost reduction) will also lead to more sustainable solutions. Driving less kilometres because routes have been optimised by changing the order in which customers are visited or changing the assignment of customers to routes will directly lead to reduced emission levels. So not Green or Lean but Lean & Green!

Distinguishing the Good from the Bad

2011-09-25T10:53:00.000-07:00

In my work to assist companies in improving their decisions making, adding mathematical rigour and making it fact based, sooner or later my client remarks that now that Operations Research is used the quality of decisions must have improved. It’s tempting to confirm that, but that would be too single minded. While using Operations Research will have a positive influence on decision quality, it is only one of many factors in high quality decision making. In judging the quality of a decision we typically equate decision quality with the attractiveness of the result. Don’t you feel silly when you’ve carried around your umbrella all day but there wasn’t a drop of rain to shield you from? What does it say about the quality of the decision you made that morning? Is good or bad determined by the result? And what does it tell you about the added value of the decision methodology you used in making the decision?

When we have a good result we are inclined to conclude that we’ve made a good decision. Likewise, with a bad result, we conclude that we’ve made a bad decision. This is definitely not true. Decisions and results are two different things. Good results are what we desire, whereas good decisions are what we can do to maximise the likelihood of good results. For decisions that are made at a high frequency (say every day/hour) quality could be measured using statistics, improving consecutive decisions. The conditions under which operational decisions lead to a result can only change slightly, given the short time span between the two. But for decisions on the tactic or strategic level it can take months or even years before achieving a result, for example in developing a new product. Using statistics to measure decision quality in that case is unrealistic. Moreover many of this kind of decisions are of the one-of-a-kind nature. When the time between decision and result increases, uncertainty will have a growing impact on the quality of the result. In the future, events can happen that cannot be controlled or foreseen. Such events can cause good decisions to have a bad result and vice versa. Therefore, the quality of the result is not a good indicator of decision quality and the result is irrelevant as a measure of decision quality.

How to assure good decisions then? Key in making a good decision is to have a structured decision making process. A structured decision making process starts with three ingredients:

What do I know (Information) about the business opportunity under consideration and the environment in which it resides?
What are the options (Alternatives) open to me?
What are my preferences (Values) in deciding between the alternatives?

Central in a structured decision making process is the logic or mathematical model. It allows you to put Information, Alternatives and Values together in a logically consistent way and make a good decision.

Because the inputs of the decision are made explicit we can establish that a good decision has been made, before the results of the decision are known. It allows discussion on all the inputs, therefore building a common view and commitment, supporting the implementation of the decision. Notice that the logic follows from all three factors, Information, Alternatives and Values. So Logic alone is not a guarantee to quality decisions. Putting this process to work starts with framing the decision; making sure that purpose and scope of the decision is discussed and agreed upon. Next is identification of what can change and can’t be changed in making the decision, creating an explicit or implicit set of alternatives. As Michael Trick blogs, accurate data is essential in achieving high quality decisions. Without it, decisions are based on quicksand. It’s the third ingredient in the decision making process. Final step before preparing a mathematical model is deciding on the valuation principles to be used. A decision is made because it will lead to an increase in value within an organisation, like increase in share price, revenue or EBIT. So valuation needs to be explicitly considered in decision making. With all these ingredients in place and managed right, good decisions will results.

So next time when you return home soaking wet because you left your umbrella at home given the weather forecaster was absolutely sure that it was going to be a sunny day, go back and review your decision making process that morning. Check the information base, the alternatives considered, the values and logic used before you consider yourself a fool. Chances are that it wasn’t a bad decision. The result was bad, but that’s because you can’t trust a weather forecaster. Some things can’t be changed; it’s your decision to prepare for them or not that makes your decision good or bad.

Complexity Defied

2011-08-15T09:01:00.000-07:00

A recent survey of KPMG among senior executives around the globe (Confronting Complexity, 2011) shows that the ability to manage today’s complex business issues is seen as one of the key factors for success. Complexity in business has increased over the past years because of changes in economic, regulatory, political and social environments. Also its causes change as companies move through the business cycle and as economies develop. Increased complexity leads to cost increases and the need for new skills within an organisation. Besides being an important challenge, the senior executives find that increased complexity also creates new opportunities, including gaining a competitive advantage and improving efficiencies. Interesting result from the survey is that technology is a critical issue, both as a cause of complexity and a key solution. New technology changes business models, enables process improvements and opens new markets, but also creates new challenges like how to incorporate it into every day business. Operations Research is one of those new technologies. It requires effort to incorporate it into the decision making DNA of your organisation, but when available it lets you defy complexity.

Operations Research has proved to be the best answer to handle complexity many times. A well known example is the way in which American Airlines used Operations Research to turn the effects of the Airline Deregulation act into an opportunity, changing the way the airline industry operated completely. Comparable to this, the Dutch based Sundio Group reinvented the online travel business with the application of Operations Research. Key for success at Sundio is to offer the best price for flight seats, hotel rooms and package trips. Finding the best price is complex, because Sundio must buy capacity at hotels, resorts and airlines before the can sell it to their customers. Because of the uncertainty in demand and the great amount of products in their portfolio, deciding on the best possible price mix is complex. A dedicated decision support system with optimisation algorithms was build to support Sundio in handling this complexity and turned it into a competitive edge.

Recently Midwest ISO, together with Paragon decision technology, won the Edelman award for their achievements in handling the complexity of the production and transportation of electric power. Midwest is responsible for the delivery of electric power across 14 US states and the Canadian province of Manitoba. In doing so it has to manage over a 1.000 power plants and nearly 60.000 miles of high voltage transportation lines to deliver electric power to 40 million end user customers. According to Midwest, the network is the most complex machine ever created by man. Due to changes in regulation, requiring open access to network transmission lines, Midwest transformed the electric utility industry in the Midwestern United States through the development and implementation of energy and ancillary services markets. Since electric power can’t be stored, Midwest needs to carefully balance supply and demand of electric power each moment in time. Changes in demand must be dealt with immediately by adapting the supply of electric power within the technical capabilities of the whole system. In order to do so Midwest has to solve a dazzling puzzle with millions of decision variables each 5 minutes. With the decision power of Operations Research, made available to decision makers across the Midwest network via the use of AIMMS, Midwest was able to achieve that. As a result Midwest ISO adds significant value to the region through improved reliability and increased efficiencies of the region’s power plants and transmission assets. It has been estimated that Midwest realized a cumulative saving from 2007 through 2010 of at least $2.1 billion.

So in confronting complexity, senior management need not despair. As the KPMG survey concludes, technology is a hot-spot. With the decision power of Operations Research available, senior management has the possibility and capability to improve or even change business models, open up new markets and electrify business performance, defying complexity.

Modelling Magic

2011-06-30T11:26:00.000-07:00

In his third law Arthur C. Clark states that any sufficiently advanced technology is indistinguishable from magic. So, in some cases at least, magic doesn’t derive from an actual mystical or spiritual source; rather, it is technology in disguise. When we would be able to send a car a couple of ages back in time, driving our horseless carriages probably would have created a witch hunt. Not to mention flying an aircraft. Time travelling over shorter distances; our computer networks and phones that can do nearly everything must seem magic for someone living in the late 1960’s. So Arthur C. Clark is probably right. What does it imply for Operations Research? Is it sufficiently advanced technology to be indistinguishable from magic? Initiated only a few decades ago in World War II, Operations Research has enabled us to achieve some remarkable things that to the non-initiated seem to be magic. It has enabled us to solve the unsolvable and step by step it is entering our daily lives becoming an indispensable piece of technology, doing its magic every day.

The simple task of assigning activities to people can become very complex and unsolvable for a human. For small instances it’s a simple puzzle. In case of 2 activities and 2 people there are only 2 possible assignments. Deciding which one is best is therefore easy. But when the size of the instance increases, the number of possible assignments explodes which makes it impossible to find the best possible assignment. In optimising the assignment of 70 activities to 70 people, there are 70! different assignments to be evaluated. Constructing all of the 1.19785717 × 10¹⁰⁰ assignment possibilities would take forever. How to find the best possible one? Here is where Operations Research does its magic. One of the first optimisation techniques developed within Operations Research is the simplex method. It was discovered by George Dantzig in the late 1940’s and it still plays a very central role in Operations Research. With the simplex method the best possible assignment of 70 jobs to 70 people can be found within a few minutes or even seconds. Compared to what a human can do, this for sure is magic.

When we fast forward to our current era, Operations Research applications arise everywhere; in Finance & Accounting, Marketing, Procurement, Production management, Logistics, Personnel management, Government, Sports, etc, etc. It has become a deciding factor for companies to survive or become top players in their market. One of the areas where Operations Research has become a deciding factor is in the airline industry. In 1978 the Airline Deregulation act was signed, its purpose was to remove government control over commercial aviation. As a consequence competition increased with low-cost carriers seizing their opportunity to get a share of the commercial aviation market. The exposure to competition led to heavy losses for a number of carriers, some of them even went bankrupt. To counter low-cost airlines like People Express, several strategies were developed by the main carriers. American Airlines (AA) was the most successful one; using Operations Research they developed a pricing strategy now known as Revenue Management or Yield Management. According to AA it is “the single most important technical development in transportation management”. The essence of Yield Management is to use price to influence customer demand. As a consequence the price for a passenger seat may vary over time. This explains why the price for the same flight may vary between two visits to a booking site, leaving you clueless on the reason why. As if the airline uses a spell to maximise revenue. But because of this magic, we can fly cheap.

Operations Research is entering our daily lives more and more. It is responsible for the fact that we:

have highly reliable electricity and gas supply networks,
can use the internet,
are able to develop reliable public transport schedules,
have attractive soccer match schedules
are able to operate global supply chains,
have satellite navigation,
can manage the risks of pension funds,
have effective cancer treatments
etc

It’s all possible because of Operations Research. It doesn’t matter if you understand why it is possible, just use it. Sit back and enjoy, let it do its magic!

Solving a Mayor’s dilemma; fact based environmental policy development

2011-06-26T11:37:00.000-07:00

This week Rotterdam, one of the major cities of the Netherlands, was on the news. The city’s Court of Audit had analysed the city counsel’s policies on improving the environmental conditions of the inner-city and concluded that things weren’t going great. Rotterdam has set itself emission reduction goals, but with the selected policies it wasn’t going to achieve them according to the Court of Audit. The Court even concluded that the situation was getting worse. Compared to 2009 the number of hot spots on NO₂-emissions in the city has increased from 1 to 5 making the city a less healthy place to be in. The city council however is still convinced that they will achieve the goals set for 2015. Until now however, several countermeasures have been introduced with little or no effect on the air quality in the inner-city. One very drastic way to accomplish emission reductions would be to close down several parts of the inner-city; this would however impact the local economy heavily. Quite a dilemma for Mayor Aboutaleb.

Major part of the air pollution in inner-cities comes from the vehicles doing last mile distribution. Think of the replenishment of shops and stores but also of bars and restaurants. Many of them use a Just-in-time (JIT) based replenishment strategy, increasing the number of vehicle visits while reducing the drop size. JIT minimises the inventory levels in the shop and hence reduces costs. Counter side of it is that the inner cities become congested with vehicles and the emission levels rise. So, reducing the number of vehicles and/or the total distance travelled by those vehicles will lower emission levels and therefore improve air quality. Easy, thing is how to set a reasonable enough level of emission reduction to improve living conditions in the inner-city without harming economic activity and second decide on the necessary measures to achieve it.

It has been the subject of a project that I have been working on in the past months. Objective of the project was to provide decision support to the municipality of cities in deciding on logistical measures to better regulate inner-city logistics and reduce emission levels. One of the objectives of the project was to supply insights on the current sustainability of a city and assist in setting reasonable emission reduction goals. The project resulted in a mathematical model that supports these decisions, fact based. Also the model is capable of evaluating different measures (like a curfew, consolidated transport from city depot, electric vehicles, etc) on congestion and emission reduction effects. That way the most effective measure can be selected before putting it in practice.

Setting a realistic goal is difficult. Cities have different infrastructures and different distribution of shops, bars and restaurants. Taking all this into account would take a lot of modelling effort and detailed information on the route the vehicles take. Information that is not available (yes, even in these data rich times, we sometimes lack the data!). Therefore we decided to use a technique that doesn’t require a lot of detail on the “production” of emissions but would give a good estimate on the “environmental performance”. To achieve this we used Data Envelopment Analysis (DEA). DEA is a technique that is used to calculate and compare the efficiency of decision making units (DMU’s), for example factories, that makes no assumption on the underlying form of the production function. It focuses on the efficiency of the transition of inputs to outputs only. Differences in efficiency tell you how improvements can be achieved. In case of the environmental performance we looked at how vehicle movements and emissions were “produced” by number of deliveries and number of drop-off locations (shops, bars, etc) in the inner-city. The resulting relative environmental performance of the cities in the benchmark tells them how well they are doing compared to other cities and how much the emission levels should be lowered to be as “green” as the most efficient city in the benchmark.

When comparing Rotterdam with Amsterdam, Utrecht, The Hague, Enschede and Tilburg I found that Tilburg is the most environmental efficient city. This is no surprise, Tilburg has been working hard to improve inner-city environment for some years already, with all kinds of initiatives. So in reviewing and perhaps setting new goals for 2015, the city counsel of Rotterdam should have a talk with Tilburg, discussing the measures Tilburg has in place to improve the environmental conditions of the inner-city. Knowing what to aim for, the Mayor of Rotterdam can than make a fact-based trade of between the countermeasures to take to improve the inner-city environment and the economic impact of them. And Mayor Aboutaleb, …..010 is doing better than 020!

Analytics and Operations Research; a practitioner’s view

2011-05-22T09:46:00.000-07:00

The topic whether the O.R. society should embrace (business) analytics is one that will probably go on for a while. It’s THE theme that keeps O.R societies occupied at this moment; all are busy with the question whether they should hook on to analytics since it could boost awareness and interest for O.R. Although the term “Business Analytics” is quite old, it dates back to Frederick Taylor’s time management practice in the late 19th century, it is presented as the latest trend in management and every CEO should start using it. Amazon lists over 1,500 books on the subject, nearly every one of these books has the theme “Start using Analytics in decision making, otherwise you will be doomed to the lower end of the performance ladder and go bust”. In promoting the use of analytics different terms are used which makes it hard to understand what people are really talking about. Terms like business intelligence, business analytics, descriptive analytics, predictive analytics, prescriptive analytics, and so on. From a practitioner’s point of view the discussion on the subject is rather academic, my clients don’t really care whether I use the term Analytics or O.R. in improving their operations or decision making. They just want me to help them solve their problem.

If have been working in O.R. consulting for over 20 years now and have learned something that Plato already knew over 2300 year ago, a good decision is based on knowledge not on numbers. It isn’t the analyses of data (=Analytics?) or building and solving a math models (=O.R.?) that leads to better decisions, it’s the knowledge gained in the process. It starts with understanding the problem and framing it right. This can best be achieved by gathering and analysing relevant data, measuring performance and identifying the applicable business rules. Analytics if you will. This analysis will increase the knowledge about the problem at hand and the environment in which it needs to be solved. Based on the data analysis and the identified business rules, directions for improvement (scenarios) can be identified. By analyzing the scenarios, the impact (consequences) of each of these can be identified, again increasing the knowledge about the problem, but also on how to solve it. The “do’s and dont’s” have come forward at this point. Next step is to use the knowledge about the challenge, the data and the business rules to build and use/solve a math model to find the best possible and achievable solution to the challenge (note: optimality in practice is something different compared to the textbook concept of optimality). With the knowledge gained during each of the above steps, implementing the solution is straight forward, apart from the “normal” potential change management issues. Result of it all is a solution to a practical challenge, and hopefully a satisfied customer.

My clients have never asked me what techniques I use to help solve the challenge they face and I also never tell them. In the past 20 years (See: Does O.R. Sell?) I have never come across a client that hired me because I could analyse data, build a forecast model, build/solve a linear programme or was able to build a simulation model. There is a simple reason for that, they don’t know the difference and they don't need to. Introducing yourself with that you are really good at building a math model, have been in Monte Carlo simulation or Markov chains for years, doesn’t help build your credibility. Talking about O.R. or Analytics doesn’t either. What counts is that you understand or show that you’re able to understand the business of your client, his organisation and the challenge he faces. So discussing whether Analytics and O.R. are the same, part of each other or complementary doesn’t really matter from my point of view. I’ll use the technique that is required to solve my clients challenge, no matter if it’s descriptive, predictive or prescriptive. Whoever thought of the term “Prescriptive Analytics” by the way? It makes O.R. to something that can only be applied when a specialist tells you how and when to use it. “Solve this LP model 3 times a day and your problem is solved?”

I once used just a blank sheet of paper to solve a business challenge, a real back of the napkin situation. One drawing was enough to identify and solve the business issue. The drawing was a simple graph. The shortest path in the graph was the solution to the client’s challenge. Calculations where not required, even my client could see the solution immediately. This shows that O.R. can be down to earth and within reach of everybody. That is also how we should go about in the Analytics vs Operations Research discussion, down to earth and for everybody including clients. I would suggest a small twist and add some special focus on the practical side of it all.