摘要
针对细菌觅食优化算法寻优速度慢且易陷入局部最优等缺陷,提出一种双菌群细菌觅食优化算法.引入菌群密度函数因子,并添加当前趋化周期内的最优细菌对其他细菌在寻优方向上进行指导,同时改进固定步长为自适应变化的趋化步长,避免了在最优解附近出现震荡现象及算法陷入局部最优;保留精英细菌的同时提出交叉算子和变异算子,有目的地在搜索区域寻找最优解,帮助早熟细菌跳出局部最优,一定程度上避免了算法早熟;提出双菌群优化机制,增加了菌群的多样性,提高了算法的全局搜索能力,有效抑制算法退化现象.对10个经典测试函数的仿真结果表明,所提出的算法与细菌觅食优化(bacterial foraging optimization,BFO)算法相比,最优解的精度普遍提高了几个数量级,且迭代次数更少,优化速度与全局收敛能力均有所提升.
A double flora bacterial foraging optimization algorithm is presented to solve the problems of slow convergence and local optimization. The bacteria density factor is introduced, and the optimization direction to another bacterium is guided by the current optimal bacteria. The adaptive chemotactic step is used instead of fixed step to avoid turbulence near the optimal solution and to obtain partial optimal solutions. The crossover and the mutation operators are proposed and the elite bacterium is retained to find the optimal solution in the search area effi- ciently. Therefore, the premature bacteria are helped to jump out of the local optimal solution to a certain extent. A double flora optimization mechanism is formulated to increase the diversity of flora, to enhance the global search capability and suppress the degeneracy phenomenon. The simulation results of ten benchmark functions have demonstrated that the solution accuracy of proposed algorithm is generally improved by several orders of magnitude in comparison with standard bacterial foraging optimization, and fewer iterations are needed. Both optimization speed and global convergence ability are improved.
出处
《深圳大学学报（理工版）》
EI
CAS
北大核心
2014年第1期43-51,共9页
Journal of Shenzhen University(Science and Engineering)
基金
国防基础科研计划资助项目(A11***007)~~
关键词
人工智能
细菌觅食优化算法
双菌群
局部最优
自适应步长
交叉算子
变异算子
artificial intelligence
bacterial foraging optimization algorithm
double flora
local optimization
adap-tive step
crossover
mutation
