Convergence Issues with Particle Swarm Optimization

[SimonBlanke]

Following up on issue #84:

PSO fails to fully converge on simple convex problems.

Expected Behavior PSO should converge effectively on simple convex optimization problems like the sphere function.

Actual Behavior PSO fails to converge properly. PSO comes close the the global optimum but fails to fully converge.

Test Case

from gradient_free_optimizers import ParticleSwarmOptimizer

opt = ParticleSwarmOptimizer(search_space, population=20)
opt.search(sphere_function, n_iter=10000)
print(f"PSO Best score: {opt.best_score}")