自适应免疫遗传算法

Adaptive Immunity Genetic Algorithm

遗传算法(GA)是基于自然遗传规则随机搜索技术的一种进化算法,但是随着实际结构的大型化和复杂化,它往往出现过早收敛的现象。在研究了算法的编码方式、控制参数和算子操作之后,就其全局收敛性的不足,提出动态自适应策略以改进其性能,在基本遗传算子的基础上,采用了免疫遗传算子和保优策略。其中免疫算子可以防止交叉变异中的个体退化,自适应策略则保持了种群的多样性,以此保证遗传算法尽快收敛到全局最优解,称之为自适应免疫遗传算法(AIGA)。随后以经典的十杆桁架结构优化问题作为例子说明算法的优越性,结果表明AIGA在随机结构优化中计算有效、结果可靠。

Genetic Algorithm (GA) is a part of evolutionary computation techniques, where the stochastic search is carried out based on principles of natural genetics, GA usually converges prematurely for larger or complex structures. Studied the coding, control parameters, and arithmetic operators, dynamic adaptive strategy is introduced to improve the capability of GA. Based on the simple genetic operators, immunity operator and elitist selection strategy are adopted. Immunity operator enables to prevent individual degenerate in crossover or mutation, and adaptive strategy to keep population diversity, which ensure to obtain the global optimal solution, so Adaptive Immunity Genetic Algorithm (AIGA) is thus named. A reliability-based structural optimization of the classical 10-bar truss problem is taken as an example to illustrate the predominance of this algorithm.