摘要
针对高维优化问题,随机初始化的粒子群算法中不同维的收敛情况不同,常用惯性权重不能很好地平衡全局搜索和局部搜索,且算法也易陷入局部最优。本文提出一种基于惯性权重维正弦调整和t分布维变异的粒子群优化算法,兼顾各维的收敛情况,较好地保持了种群的多样性。通过4个典型函数的测试,结果表明改进算法提高了收敛速度和精度。
For high dimensional optimization problems, convergence of different dimension is different to random initialization of particle swarm optimization algorithm, the commonly used inertia weight can not balance the global search and local search very well and the algorithm is easy to fall into local optimum. This article introduces a new particle swarm optimization algorithm based on dimension sine changing inertia weight and dimension mutation in line with t distribution which taking convergence of different dimension into account and keeping the population diversity very well. Experimental studies are carried out on four classical func- tions, and the computational results show that the algorithm has faster convergence speed and higher convergence accuracy and rate than traditional article swarm optimization algorithm.
出处
《计算机与现代化》
2014年第6期84-87,93,共5页
Computer and Modernization
关键词
粒子群优化
维正弦调整
T分布
维变异
多样性
particle swarm optimization
dimension sine changing
t distribution
dimension mutation
diversity