自适应细菌觅食算法在优化问题中的应用

细菌觅食算法在求解优化问题时,以固定的步长进行趋向操作,同时以固定概率对细菌个体进行随机驱散操作,虽然可以一定程度上增加种群多样性,但是在进化后期容易使优秀的个体流失,影响算法的寻优质量.针对上述问题,论文提出步长自适应调整和驱散概率自适应调整两项改进策略,分别根据算法进化程度和细菌个体的能量值动态调整趋向操作的步长和驱散操作的概率,从而使算法在保证种群多样性的前提下,保持细菌个体具有较高觅食能力,促进算法局部搜索和全局优化的平衡.对标准测试函数和TSP问题的测试结果表明:基于自学习的细菌觅食算法具有较强的全局寻优能力,适合求解高维复杂优化问题.

Self-adaptive bacterial foraging algorithm and its application to optimization problems

The search procedure was performed in a fixed step length and the random dispersal was processed in a fixed frequency when applying bacterial foraging algorithm in optimization problems .Diversity of population could be increased through the traditional algorithm , however ,efficiency of the algorithm was compromised for possible loss of samples in later period . Two improvements , including the self‐adaption of the search step length and dispersal probability according to the evolution level and the energy value of bacteria ,were proposed in this study . The improvements were mainly aimed at enhancing the balance between local search and global optimization while keeping the diversity of populations .The proposed algorithm had been applied on the standard test function and TSP problem ,and the results indicated that the algorithm was suited for complex high dimensional optimization problems with better global search capability .