摘要
针对标准微粒群优化算法微粒运动轨迹的收敛性进行了分析,给出并证明了微粒运动轨迹收敛的充分条件.提出一种简便的等高线图判别法,该方法能够通过参数的位置判断微粒轨迹是否收敛并衡量收敛速度.为提高算法的收敛速度,构造出一种梯度微粒群优化算法,给出并证明了该方法收敛的充分条件.仿真结果表明,梯度微粒群优化算法具有优良的搜索性能.
The convergence of standard particle swarm optimization algorithm is studied. The sufficient condition for the convergence of the algorithm is given and proved. And a kind of convenient contour map discriminance is proposed. This discriminance can be used to judge if the algorithm is convergent and measure the convergence rate. A sort of gradient particle swarm optimization algorithm is presented to enhance the convergence rate of the algorithm. The sufficient condition for the convergence of this method is given and proved. Simulation results verify the correctness and efficiency of the proposed method.
出处
《控制与决策》
EI
CSCD
北大核心
2009年第4期560-564,共5页
Control and Decision
基金
上海市教委重点学科建设项目(J50602)
上海市教委科研创新项目(08ZZ78)
关键词
微粒群优化
收敛性
梯度
Particle swarm optimization
Convergence
Gradient
