# Is there software to solve this problem?

1. Jan 3, 2012

### danielci

Hi all,
I have this challenge and I am not a math guy. Can you recommend a software to solve this kind of problem?
Apologies upfront, I am not mathematically savvy (and English is my 2nd language) and for sure there could be better ways to explain this.

You have 4 variable numbers: W,X,Y,Z
You have a serie of these numbers with 100 (400 numbers total) of them: W1to100, X1to100,Y1to100, Z1to100
For each serie of numbers (Wn,Xn,Yn,Zn) you have a multiplier (A,B,C,D) affecting the whole serie
You need to compare the multiplied numbers among the others like in this example:

Example:
W1*A>X1*B = TRUE
X1*B>Y1*C = FALSE
Z1*D>W1*A = TRUE

Once this calculation is made for the "n" (100) cases of W,X,Y,Z and the probability of the result is calculated (For the example: Probability of obtaining "True, False, True" in the rest of the serie)
you need to optimize the multipliers (A,B,C,D) to obtain the best case in terms of probability.

Please provide your suggestions regarding what software can solve this kind of optimization problem. I tried Excel already, it works for the calculation but not for the optimization unless done manually for each case of A,B,C,D

Thanks!
danielci

2. Jan 4, 2012

### Bill Simpson

Excel optimization
you can find lots of information on how to optimize in Excel.

Your problem seems similar to neural network training. There are dedicated software programs for this and other examples showing how to implement these in general purpose languages.

If you could provide information on what languages you already know, if any, outside Excel then that might help someone recommend a tool for you.

You might also provide a smaller example, perhaps 4x4 numbers with reasonable equations and show the process you want to go through. That is small enough to be understandable and still possible to work through to see exactly what you are asking for.

3. Jan 4, 2012

### danielci

Bill, thank you for the guidance. I am an old time programmer who is a little bit rusty. I will follow your advice about Excel optimization. Also, a friend recommend me to try Lisp.
Regards
danielci