邻域学习细菌觅食优化算法

An Improved Bacterial Foraging Optimization Algorithmwith Neighborhood Learning

针对细菌觅食算法在求解复杂问题时优化性能不佳的缺陷,本文抛弃了传统算法改进时,参数改进及混合策略的思路,从拓扑结构的视角,首先分析了邻域拓扑结构对算法产生的不同影响;其次,借鉴PSO算法中的邻域学习策略,结合BFO算法局部搜索能力较强的特点,提出了一种邻域学习细菌觅食优化算法(BFO-NL)。最后,为了评估新算法的优化性能,将NFO-NL算法与原始BFO算法进行比较,仿真实验在8个标准测试函数中完成,结果显示BFO-NL算法收敛速度更快、求解精度更高。BFO-NL算法与原始BFO算法相比较,新算法的求解精度更高。

The original BFO algorithm shows poor performance when solving complex problems.In this paper the traditional improvements of algorithm were abandoned,such as adjusting of the parameters and hybrid strategy and ideas,but from the perspective of topology,firstly the different effect was analyzed,generated by the neighborhood topology;Secondly,by taking the neighborhood learning strategy of PSO as an example,combining with a strong local search capability of BFO,a bacterial foraging optimization with neighborhood learning(BFO-NL)is proposed.Finally,to evaluate the performance of the new algorithm,the NFO-NL algorithm is compared with the original BFO algorithm.Simulation results from eight benchmark functions show that BFO-NL algorithm converges faster and gets higher solving accuracy.Compared with the original BFO,BFO-NL algorithm can make solving accuracy higher and is more robust.