传统的粒子群优化算法(Particle Swarm Optimization,PSO)只考虑了最优粒子对整个进化过程的引导作用且在一次迭代中所有粒子采用相同的惯性权值。为了体现各粒子相对于已知最优解的差异,提出了一种基于距离度量的自适应(k,l)PSO算法。(k,l)PSO算法采用轮盘赌策略在k个最优的粒子中选择一个粒子作为全局最优粒子参与粒子的速度更新,同时,根据粒子间的平均距离l确定粒子与选中的最优粒子的距离,自适应调整粒子的惯性权值。通过基准测试函数对算法进行了实验,实验验证了(k,l)PSO算法的有效性。
The classical Particle Swarm Optimization(PSO) neglects the difference among particles and uses a fixed inertia weight in one generation.To cope with this issue,a novel method called(k,l) PSO is proposed in this paper.The(k,l) PSO chooses one of the top k particles as the global best particle according to the roulette strategy and tunes the inertia weight value according to the distance between the current particle and the global best particle.Several classical benchmark functions are used to evaluate the(k,l) PSO.The experiments demonstrate the efficiency and effectiveness of the proposed(k,l) PSO.