MILP using Genetic algorithm - penalty value increases beyond the global minimum

I have been using GA to solve a minimization problem and this involves interger variables as well. My programm is working fine except for one abnormality observed in the penalty value vs generations graph.

As you cna see here, (Best value in Black and mean value in blue) best values jumps from around 0 to 1.3. Has anyone come across a similar issue and how you could oivercome it.

Thanks.

Answer 1

[Dheeraj Singh]

You can refer to the following example in the link to know more about solving MILP using GA:

https://in.mathworks.com/help/gads/examples/solving-a-mixed-integer-engineering-design-problem-using-the-genetic-algorithm.html

Alternatively, if you are trying to solve MILP it would be more robust to use intlinprog. Also, you can verify your results you get from GA algorithm and see if there is any abnormality.

You can refer to the following link to know more about solving MILP using GA:

https://in.mathworks.com/help/gads/examples/solving-a-mixed-integer-engineering-design-problem-using-the-genetic-algorithm.html

[Udayanka Mulleriyawage]

Thank you so much for the response. I used intlinprog to solve the problem I had and then used cplex for matlab. The solver cplexmilp appraches to the optimal solution much faster than intlinprog.