一种基于分层模糊控制的免疫遗传优化算法

针对标准遗传算法的不足,借鉴生物免疫机理和人脑模糊思维功能提出一种新的基于分层模糊控制的免疫遗传算法.该算法利用免疫系统独特性网络学说,改进标准遗传算法选择算子,提高了种群多样性;同时从环境、种群、个体和基因角度,全面分析算法寻优性能和各种进化参数的启发式模糊关系,采用模糊推理动态调整交叉率、交叉位置和变异率,减小了标准遗传操作的随机性.实验结果表明,新算法不仅可有效克服标准遗传算法的缺陷,而且收敛速度、计算精度和算法稳定性也得到明显提高.

Immune genetic optimization algorithm based on multilayer fuzzy control

Aiming at the insufficiencies of standard genetic algorithm (SGA), a novel immune genetic algorithm based on multilayer fuzzy control (MFCMGA) is proposed, which adopts the biological immune theory and fuzzy thinking function of human. In order to increase the diversity of population, MFCMGA uses idiotypic immune network theory to improve the selection operator of SGA. Meanwhile in order to decrease the randomicity of SGA, the crossover probability, crossover position and mutation probability are adjusted dynamically by using fuzzy inference, which is based on analyzing the heuristic fuzzy relationship between algorithm performance and evolutionary parameters from the viewpoints of environment, population, individual and gene. Simulation results show that MFCMGA effectively overcomes the shortcomings of SGA, and evidently improves the convergent speed, computing precision and algorithm stability.