摘要
针对PSO算法所存在的早熟问题,提出了一种新的优化方法,即跳蚤算法。此算法在粒子之间加入排斥力,使得各个粒子无法聚集在同一点上,从而整个粒子群不可能趋同于局部最优解,跳蚤算法不需要假设函数最优解在粒子群运动轨迹包络体之内。采用F(x1,x2)=sin(r)/r,其中r=x21+x22(1/2)等函数验证了该算法的寻优效果。
Aiming at PSO premature problem, a new optimization method is proposed, namely flea algorithm. This algorithm adds the repelling force of particle to make each particle not be able to gather at the same point, and the entire particle swarm convergence in local optimal solution is impossible. Flea swarm algorithm does not need to assume the function optimal within the trajectory envelope body of particle swarm. Finally, the algorithm effect is verified such as F(x1,x2 ) = sin( r)/r, r =r=√x1^2+x2^2
出处
《沈阳理工大学学报》
CAS
2015年第4期80-83,共4页
Journal of Shenyang Ligong University
关键词
PSO
斥力
跳蚤算法
PSO
repelling force
flea algorithm