MATLAB Central Newsreader Recent Posts

forced discard of ODE step

Mateusz Gos — Sat, 25 May 2013 03:17:17 +0000

Hi,

I have a set of differential equations solved with ode45. At times it might be possible for the solver to find a valid time step (meeting the tolerance criteria), but for some of the variables in that time step to be physically not possible. So in other words such step is correct mathematically, but not physically.
Of course I can check for these variables, but how do I make the solver discard a particular iteration and start a new one with different time step, regardless of what it has determined tolerance-wise?

Thank you,
Matt

Re: Minimum data set size required for Kruskal-Wallis test?

Kate J. — Sat, 25 May 2013 00:18:11 +0000

>> You've concatenated three row vectors into a longer row vector. Try this
instead:
>> myData = [dataSetA' dataSetB' dataSetC'];

Adding the single quotes to format my matrix as columns did indeed work, Tom; no more error messages. Thanks!!

Re: Noncentral Chi Distribution

Josh Carmichael — Fri, 24 May 2013 21:26:08 +0000

The original message content is below. Here is my answer:

>> y = 2.*abs(x).*ncx2pdf(x.^2,v,delta);

Does the trick. Recall the transformation from one pdf to another:

Pr( Y > y) = Pr( sqrt(X) > y ) = Pr( X > y^2)

Since the cumulative distribution function is Pr( Y > y ), differentiate both sides to get:

f_Y( y ) = 2*abs(y)*f_X(y.^2)

Since the support of x is [0, Inf), the abs is implied, but this is safer. Hope that helps.

"Soenke Engelken" <soeeng@remove.prefix.gmx.de> wrote in message <ef5a701.-1@webcrossing.raydaftYaTP>...
> Dear Matlabers,
>
> I am looking for an evaluation function for the noncentral Chi
> distribution. The nonc. Chi-square is included in the statistics
> toolbox, but this one is not. Do you know of any implementation, e.g.
> using a transformation from the nonc. Chi-square? Googling or search
> on this site did not reveal any. Many thanks!
>
> Best,
> Soenke

Re: Simulink, ODE15S and stiff-system solutions getting unnecessarily stuck.

Fri, 24 May 2013 20:35:07 +0000

Mike Tocci <mtocci@mathworks.com> wrote in message <clu67s$53d$1@fred.mathworks.com>...
> Kelvin Hales wrote:
> > I find when solving my present stiff Simulink 5.1 (R13SP1) system
> > with ODE15s, that the solution tends to grind to a halt when
> > approaching a steady-state, i.e. most annoyingly at the very point
> > where it should start to take bigger steps and zoom along. If,
> > however, I stop the simulation and re-start it by initialising the
> > states with the end state of the previous run, then the solution
> > immediately picks-up and shoots along.
> >
> > Any ideas anyone?
> > Its as though the solver is getting bogged down in the detail
> > because it cannot forget the previous small time-steaps.
> > A problem with a zero-crossing detection, maybe?
> > Is there something in the Simulink debugger/profiler that would
> > reveal what's happening?
> >
> >
> > Kelvin B. Hales
> > Kelvin Hales Associates Limited
> > Consulting Control Engineers
> > Web: www.khace.com
>
> Hi Kelvin,
>
> I'm not 100% sure, but it could be zero-crossings that are causing
> this. Whenever you stop a simulation and use the final state as an
> initial state, there is some information not stored in the state
> vector that is lost. This could be affecting the zc detection in
> some blocks.
>
> I do recommend using the Simulink debugger to try to find the problem.
> If you use the command-line debugger (sldebug), there are commands to
> print out solver information. I'd try setting a time break point
> (tbreak) so you get to the point where the model nears steady state,
> and then use the command "strace 4" to print out all the information
> in the solver. If there's any information in the output that's not
> clear, let me know and I'll see if I can help. There can be alot of
> information printed to the screen, so I recommend taking a few time
> steps using "step top" to see if you can tell why the time steps are
> so small.
>
>
> Hope that helps,
>
> Mike
>

I too have the same issue. I'm simulating a highly dynamic system with lots of power setpoint jumps with ODE15s. The model matches all the extreme dynamic changes fine until a large drop in power after being held at a steady state value. At this large drop in power, the model starts to progress at a snail pace. The step size it progresses at is at about 8e-8 seconds.

I am playing around with the minimum step size, maximum step size, relative error tolerance, and absolute error tolerance, but nothing is getting me past this point.

I am attempting to use a fixed step solver, but am not partial to this because it is so slow.

I will look into the suggestion posted by Mike. Any other suggestions are appreciated.

Thanks,

J

How to customize location of startup.m?

kj — Fri, 24 May 2013 20:22:33 +0000

What's the recommended way to customize the location of the file
startup.m?

(I imagine that the answer to this question will differ depending
on the OS. If so, I'm interested in "all of them", i.e., Unix(-like)
and Windows.)

Thanks in advance!

Merge ginputc function with Display Cursor Location function

Armindo — Fri, 24 May 2013 19:58:08 +0000

I am using this very nice function ginputc
http://www.mathworks.com/matlabcentral/fileexchange/38703-custom-ginput/content/ginputc.m

That alow me to make zoom pan and other cool stuff. However i also need to be able to get the mouse position and for that there is also this function - Display Cursor Location from here http://www.mathworks.com/matlabcentral/fileexchange/20892

can anyone helpe please merging this Display Cursor Location into the ginputc function please.

best regards

Re: Inheritance from classes withina package

Steven_Lord — Fri, 24 May 2013 19:57:54 +0000

"Ron Lu" <dlu@masimo.com> wrote in message
news:knlg51$ktq$1@newscl01ah.mathworks.com...
> "Steven_Lord" <slord@mathworks.com> wrote in message
> <knlbmr$6go$1@newscl01ah.mathworks.com>...
>>
>>
>> "Ron Lu" <dlu@masimo.com> wrote in message
>> news:knjldr$1rs$1@newscl01ah.mathworks.com...
>> > "Nick Clark" wrote in message <h9dqf5$e15$1@fred.mathworks.com>...
>> >> Hi all,
>> >> I have a query regarding inheritance of one class by another within a
>> >> package. For example, classB < classA within package mypack. From
>> >> outside the package directory I type:
>> >>
>> >> import mypack.*
>> >> x = classB
>> >>
>> >> . . then I get an error saying that classA cannot be accessed. I can
>> >> resolve this by making classB < myclass.classA at the top of the
>> >> classdef of classB. However, this seems like a bit of a clunky 'duct
>> >> tape' kind of work around. It also causes problems when trying to
>> >> display the documentation about the classes which are linked in blue
>> >> from the disp command.
>> >>
>> >> What is the recommended elegant method for inheritance within a
>> >> package?
>> >
>> > Anyone can help?
>> >
>> > I ran into the exact same problem.
>> > I already finished all my classes and now want to put them into
>> > different packages. And I HAVE TO add those import statements
>> > everywhere, or add packageName.Class* to let the class know where to
>> > find dependencies in the same package.
>>
>> That is correct.
>>
>> > Is there a way to work around?
>>
>> No. Packages define a namespace, and the package name is sort of part of
>> the name of the class.
>>
>> As a concrete example, if you had this directory structure (with
>> mydirectory on the path):
>>
>> mydirectory\classB.m
>> mydirectory\+mypack\classA.m
>> mydirectory\+mypack\classB.m
>>
>> and you wanted to create an instance of the class defined in the classB.m
>> file inside +myPack from within the classA.m file, you would need to use
>> mypack.classB. If instead you wanted to instantiate an instance of the
>> mydirectory\classB.m class, you would simply use classB. [If you imported
>> mypack.classB inside classA.m, I don't believe you'd be able to access
>> the classB class defined outside the package in the function in which you
>> imported.]
>>
>> --
>> Steve Lord
>> slord@mathworks.com
>> To contact Technical Support use the Contact Us link on
>> http://www.mathworks.com
>
> Thank you Steve!
>
> I find that even when I call classA's method within classA (like one
> method uses another), I also need to use mypack.classA.method1 in
> classA.method2.

For _static_ methods, yes. The package name is sort of part of the name of
the class, and to call a static method you need <name of the class>.<name of
the static method>. There is usually no instance of the class involved in a
static method call, so MATLAB needs a different way to know which class's
method to call.

For non-static methods, no. If you call method1 with an instance of
mypack.classA as input, the method that will be called is
mypack.classA.method1. [Basically. I'm ignoring N-ary (N > 1) functions with
multiple object inputs for simplicity.]

> Besides, import* only works within a method, then every class method
> needs a import statement. That means if I redistribute my code in a
> different package name, I will need to change all those package name
> instances. Is that right?

Yes. Changing the package name in which your program files are located is a
major reorganization of your code, and will likely require changes to users
of your code (including your code itself.) Consider when a professional
sports team changes its home city. Its name changes. Its logo changes. Its
mascot may change. The equipment that has the team's old logo on it needs to
change. The merchandise (T-shirts, hats, etc.) with the team's old name and
logo need to change.

> Does Mathworks have a plan to enhance the way using packages?

I recommend that you contact Technical Support and describe the specific
enhancements you have in mind for packages and the packages workflow to the
Support staff. They can capture your use case and requests in the
enhancement database, or may be able to suggest alternatives that I haven't
thought of for ways to do what you want now.

--
Steve Lord
slord@mathworks.com
To contact Technical Support use the Contact Us link on
http://www.mathworks.com

We need a transfer function of a basic double acting pneumatic cylinder!

ömer — Fri, 24 May 2013 19:25:11 +0000

Hi Guys,

For semester project, we need transfer function of a basic double action pneumatic cylinder.

Thank for your help.

Real time matrix multiplication

Ed Yu — Fri, 24 May 2013 18:50:16 +0000

Ching-Yuih,

Search for matlabcontrol which allows you to communicate from java to matlab.

Good Luck!
Ed.

Re: Memory leak writing NetCDF4 files but not NetCDF3 files

Ed Yu — Fri, 24 May 2013 18:47:11 +0000

Nevermind Jeff... I thought you were using the java library... I guess you'll have to talk to matlab technical support for this... Ed.

"Ed Yu" wrote in message <knoc69$1b4$1@newscl01ah.mathworks.com>...
> Jeff,
>
> It may have to do with the netcdf library implementation of the version 4 file handling... Maybe upgrading the library will help?
>
> Good Luck!
> Ed.

Re: Memory leak writing NetCDF4 files but not NetCDF3 files

Ed Yu — Fri, 24 May 2013 18:40:10 +0000

Jeff,

It may have to do with the netcdf library implementation of the version 4 file handling... Maybe upgrading the library will help?

Good Luck!
Ed.

Re: integrating java and matlab

Ed Yu — Fri, 24 May 2013 17:40:07 +0000

Pubudu,

You are on the right track... There are other options such as JNI... You can search for it in the forum here...

Good Luck!
Ed.

Re: MATLAB Vs JAVA

Ed Yu — Fri, 24 May 2013 17:39:07 +0000

Lipi,

That depends on if you are mostly doing matrix operations... If so, MATLAB will be great, if not, java will be better.

Good Luck!
Ed.

Receiving image Data as Binary Array and Then Display it

Ed Yu — Fri, 24 May 2013 17:37:07 +0000

Hi Lavan,

First make sure you are receiving the file "exactly as is" because there might be platform differences as big and little edian machines. The best way is to write it out to a file and then compare if they are identical. Once you are sure they are identical, make sure you write it out with the right file extension then you can use something like the image toolbox to read it back and display it.

Good Luck!
Ed.

Re: MATLAB GUI Prototyping Approaches

Ed Yu — Fri, 24 May 2013 17:33:07 +0000

Hi Svetlin,

Guide is the way MATLAB allowing you to use an interactive tool to deisgn GUI. You can also use java UI components within MATLAB. Look at the undocumentedmatlab website and you'll have an idea what you can do...

Good Luck!
Ed.