离散粒子群算法多样性维持机制及其应用

Diversity Maintenance Mechanism for Discrete Particle Swarm Optimization and It's Application

针对标准的离散粒子群算法在进化后期种群出现'趋同性',从而造成搜索能力减弱和收敛速度缓慢的缺点,提出了一种在压缩空间中维持种群多样性的机制,保证了算法在进化中后期仍具有较强的搜索能力,提高了收敛速度和精度。将提出的算法用于组合优化问题,计算机仿真证明了该算法的有效性。

A mechanism to maintain the population diversity in compressed dimension is proposed against the weakness that similarity occurs among the population in later period of evolution in standard discrete particle swarm optimization so that the searching capability lessens and convergence rate slows down, therefore it has ensured the effective searching capability of the swarm optimization in later period of evolution to improve the convergence rate and precision. The proposed algorithm is applied to combination optimization problem, and its validity is verified by computer simulation.