Dear Eric,

By mentioning 'swarm' I assume that you are talking about the input variables. Kindly check if a condition is mentioned in your code which forces the variable to take upper bound or lower bound value if they exceed the specified upper or lower bound respectively.

Hey Abhinav, thank you for your response. I am generating a swarm in PSO, my upper bound is positive and my lower bound is zero. I multiply my upper bound value by random numbers between 1 and zero. I do not understand how some turn out to be negative

follow PSO example by Rao, in the book:

Engineering Optimization

chapter 13, Example 13.4

Dear Eric,

To comment exactly what is wrong with your methodology, I need to see the problem and its code. But As mentioned by Dr. Ibraheem, you look at the example mentioned by Rao in Engineering Optimization book. You solve same example through your PSO implementation step by step. Hopefully you may discover what exactly is wrong in your implementation.

Thank You Abhinav and Ibraheem

hope you overcame this issue

Hey Ibraheem, I am Trying.Actually, I have used all the values mentioned in the book. Maybe you give me your email I forward the whole code. I am doing load flow using BFSM. Since I am using using IEEE 33 bus test system, I do not know whether the issue is from placing the DG at a particular bus wrongly.

Dear Erick,

If I understand your problem and assuming there is no problem with your code,

moving particles at each iteration don't prevent them to get out the definition space of each variable.

You need to check if such a thing is happening and choose a behaviour strategy such as stop at the frontier, rebound into the search space, ...

You can find more about this in litterature such as "Confinements and Biases in Particle Swarm Optimisation" by Maurice Clerc.

Good luck !

Thanks Peio Loubiere

Erick Go, how could you solve the problem of getting negative values of PSO? i couldn't find any solution around.