FST-PSO’s optional settings

Marco S. Nobile, Associate Professor with the Ca' Foscari University of Venice, Italy

The following arguments are supported:

  • you can pass the number of iterations as argument (max_iter) to solve_with_fstpso(). Default is 100 iterations;
  • you can use a non-uniform initialization of the particles using the optional argument creation_method of solve_with_pso(). For example: solve_with_fstpso(creation_method={'name': 'logarithmic'});
  • you can also override the heuristics for the swarm size, by explicitly setting the number of particles with the method set_swarm_size().

You can disable the fuzzy rules for the social factor, cognitive factor, inertia weight, minimum velocity, and maximum velocity by using the following methods before calling solve_with_fstpso():

  • disable_fuzzyrule_social();
  • disable_fuzzyrule_cognitive();
  • disable_fuzzyrule_inertia();
  • disable_fuzzyrule_minvelocity();
  • disable_fuzzyrule_maxvelocity().

You can provide a list/array of guesses to the initial population using the optional argument initial_guess_list of the solve_with_fstpso() method. Please note that the number of guesses must be smaller than the swarm size.