自适应多策略花朵授粉算法

Self-adaptive flower pollination algorithm with multiple strategies

针对花朵授粉算法极易陷入局部最优解且寻优精度不高的问题,提出自适应多策略花朵授粉算法(self-adaptive flower pollination algorithm with multiple strategies,SMFPA)。利用锚点策略提高种群的多样性,采用摄动策略改善全局勘探能力,采用局部搜索增强策略提升其开采最优解的能力。为验证SMFPA的性能,比较5种算法在解决12个测试问题上的寻优结果,实验结果表明,在寻优速度以及寻优精度方面,SMFPA算法表现更优。通过比较算法在管柱设计问题上的寻优结果,进一步评估SMFPA的寻优性能。

To solve the problem of being easy to plunge into local extremum and the low accuracy computation in basic flower pollination algorithm,a self-adaptive flower pollination algorithm with multiple strategies(SMFPA)was presented.An anchor point strategy was used to enhance the diversity of population.Perturbation strategy was introduced to improve the global exploration capability,and a local search enhanced strategy was adopted to improve the ability of exploiting optimal solution.To verify the performance of SMFPA,the optimization results of five algorithms on twelve test functions were compared.Experimental results show that SMFPA performs better in convergence speed and optimization accuracy.The optimization performance of SMFPA was further evaluated by comparing the optimization results of the algorithm in the tubular column design problem.