A sample of Particle Swarm Optimization(PSO) for Traveling Sales-person Problem(TSP)
Source Code(Ruby)
Trajectory of agents(Parameter C=1.0)
Trajectory of agents(Movie)(Parameter C=1.0)
Trajectory of agents(Parameter C=2.0)
Trajectory of agents(Movie)(Parameter C=2.0)
The evolution of optimization
where the parameter C is velosity of agents. In the case of C=1.0 , the convergence is relatively fast, but it goes to a local minimum. C=2.0 finds more optimal answer.
Fitness Landscape(The lowest part is the optimal answer)
Space: 120 x 120 2D integral space. A position means 2 answers for independent Traveling Sales-person Problems.
Reference: