why is my code wrong ?

Please see the files attached and let me know what is wrong with this

l_dp =

    0.7101         0         0    0.0447    0.7701         0    0.0379    0.8407
         0         0         0    1.5729    0.0635         0         0    0.5264
    1.1636    0.4907         0         0    0.2497         0    1.0603         0
         0         0    1.1273         0    0.1478         0         0    1.4100

and this not right since demand on a path should total to one ie the sum of each row should be 1

whats wrong with my code is it ub limit or what ?

Please do help

[Geoff Hayes]

Bhavz - where in your code do you force that condition (that each row should sum to one) to be true? Or where do you expect that to be enforced (and why)?

[Bhavz]

^ its not about enforcing the condition but shudnt the demand on a path total to one technically ... thats what the correct functioning wud be ? I am attaching the equation i want to implement constraint wise ,u can see if i have made a mistake there What I have done is taken the network optimization from paper 2 in the network from 1 and I am trying to measure the effect on link utilization. How would you change the given Algorithm to route unsplittable flows on the specified single primary path only

And as always thankyou so much for yr help

Answer 1

[Matt J]

ie the sum of each row should be 1

If the unknown vector x is supposed to satisfy this, then presumably you need to have linear equality constraints,

   Aeq=kron(ones(1,8),eye(4));
   beq=[1;1;1;1]

   [x,fval,exitflag] = ga(@objectiveFun,... 
    32,[],[],Aeq,beq,lb,[],nonlcon,options);

[Bhavz]

Many Thanks for yr answer the equations I am trying to implement is this one attached

[Matt J]

You're welcome.

[Bhavz]

^ Do you find anything wrong in my implementation given the equationsof constraints I am trying to implement

[Matt J]

The constraints in formulas.bmp are linear inequality constraints A*x(:)<=b, which are specified in the 3rd and 4th argument to ga().

   x = ga(fitnessfcn,nvars,A,b)

But it doesn't look like you've passed to any A, b data to ga.

[Bhavz]

Oh How would I do that as in pass values from constraints.m to

   [x,fval,exitflag] = ga(@objectiveFun,... 
    32,[],[],Aeq,beq,lb,[],nonlcon,options);

[Matt J]

You would not use the nonlinear constraint function to handle linear constraints. You would generate appropriate A, b matrices, similar to what I did for you with Aeq and beq.