Do you want me to do the coding for you?

not the exact coding. Just help me with how to generate random population initially and how to make sure the cobatraints are not violated?

@ Abhinav I reckon you mean constraints not cobatraints.  Did you check the computational complexity of the problem? If the problem is polynomially solvable, apply PSO to solve a particular problem. Use a commercial software package like LINGO to solve the same problem. Check the solution validity by checking the final solutions (as they are optimal, so the values should be same). If the problem has higher complexity, then the approach may be different.                                                 You may use the concept of developing PSO for production scheduling problems, like job shop, open shop or flowshop scheduling problems, as these problems have some constraints of their own and while representing a particle, these constraints must be satisfied. 

for example you have a string like 

| 0  | 1    | 0    |  

this three length string can be converted from binary to decimal. the minimum value is 0 and the maximum is 7 (111).

If you divide the resulting number to 7 then you can always have a number between 0 and 1. I mean you can scale it.

then multiply it to any other range you want.

You can always avoid violating the constraints but you can try the mentioned method as well as penalizing teh violation and adding it as a term to objective function.   

See this: http://www.sciencedirect.com/science/article/pii/S0045782501003231

http://www.sciencedirect.com/science/article/pii/S0045782599003898