I use ant colony optimization to solve a problem. In my case, at each iteration, n ants are generated from n nodes (one ant per node every iteration). I obtain solutions that verify the conditions of the problem. But, I don't achieve a convergence (for example, I have 30 iterations, the best solution is obtained in the iteration 8 or 9). I want to know if using only a single ant at each iteration is the problem? Also, I want to know if an ant colony algorithm must converge to a state of equilibrium? thank you in advance.
Asked
Active
Viewed 151 times
1 Answers
0
Convergence and divergence of heuristics algorithms is a very broad topic. Your problem type, dimension, parameters affect the behaviour of the algorithm. You should study the paper in here http://iridia.ulb.ac.be/IridiaTrSeries/rev/IridiaTr2009-013r001.pdf for basic information about ACO algorithms.
After then you should ask a question based on https://stackoverflow.com/help/mcve.
kur ag
- 591
- 8
- 19