keeps getting same values when running simulated annealing

Please I need to how I can stop getting the same random numbers of parameters when I am running the SA. Also, at a point, i get error after running a full iteration which makes it impossible to re-estimate my parameters. Why do I keep getting the same parameters during each iteration? why is the output no changing? Is there any other way of estimation?

Answer 1

[Walter Roberson]

http://www.mathworks.com/help/gads/using-simulated-annealing-from-the-command-line.html#bq2g25b-41

The default is that you would not get the same outputs each time. Are you using the tool to set the parameters? I have not looked at the options in the tool.

[Victor]

Walt, I set up the code to produce different results, re-calibrate parameters. When I did so with earlier model, I got a result but in this case, I couldn't. My target for x was to be between 0-1.6 and I'm trying to minimise it from 10.80. I have about 30 parameters that are expected to change, they only change for 3 iterations then remain the same for the remaining 100 sequence. Then error appears when a loop is about to start. 'Index matrix must agree'.

[Walter Roberson]

Please show the error traceback.

[Victor]

Index exceeds matrix dimensions. This is the error it shows plus others that points to my code. But once I change the parameters back to its original value, it starts iterating again.

[Walter Roberson]

At the command line, give the command

dbstop if caught error

and then run your program. When it stops, look through the size() of the various variables and expressions. Work backwards to figure out whether the problem is with a value of the wrong size being passed in, or due to a calculation not returning as many values as expected. If it is due to something of the wrong size being passed in, use dbup to go up a stack frame and look at the expression that was passed in that position; trace the expression back up through, using dbup as needed until you find a place where a calculation is not returning the expected size or you have gotten all the way back to the values you supplied.

For us to assist you, we would need to full trace of the error, such as the lines that indicate which line of which routine you were executing in, and what the text of that line is, and which line of which routine called it and the text of that line, and so on, right back to your code.

Answer 2

[Victor]

Like I told you, I know what causes the error. The problem I am facing is that x does not change in SA. The SA code is running another file which re-estimates my parameters, this seperate code gives different output but the x doesn't change.

[Walter Roberson]

How does x not changing get to the point of an array indexing error?

[Victor]

please stop asking me annoying questions. Thank you for your attempt to help

[Walter Roberson]

You have not shown us the code, you have not shown us the error traceback, and you appear to be declining to answer questions that might help us to debug your difficulties. Have you solved the problem now, or have you given up on it, or are you hoping that someone might read your mind?