Abstract: We proposed an improved particle swarm optimization (PS O) algorithm. The algorithm used reasonable balance between the global explorati on and local development, which reduced the possibility of premature convergence of PSO. Firstly, the Beta distribution was used to initialize population. Secon dly, the inverse incomplete Γ function was used to update the inertia weig ht. Thirdly, a new operator based on differential evolution was introduced to ad just the velocity. Finally, we used the method based on boundary symmetry ma pping to deal with the cross boundary of particles. Numerical simulation results show that the improved algorithm is obviously superior to the common PSO algori thm, differential evolution algorithm, artificial bee colony optimization algorithm and an t colony optimization algorithm.

Key words: Beta distribution functi on, algorithm design, numerical optimization, particle swarm optimization, inverse incomplete Γ function