摘要
遗传算法极难处理高维约束优化问题,故借鉴免疫系统机理,提出一种抗体修正免疫算法解决一类高维约束优化问题。该算法设计的关键在于抗体亲和力由抗体浓度及群体状态决定;可行抗体被克隆、突变;非可行抗体的基因按价值密度由小到大逐一修正。选取两种已有的智能算法(ETGA、ISGA),通过不同约束条件下的高维0/1背包问题的仿真比较。结果表明,该算法较其他算法能更快地跟踪最优值,具有较强的勘测和开采能力。
It＇ s difficulty to deal with high-dimensional optimization problem constrained for GA, this paper proposed an immune algorithm with antibody-repaired, based on biological immune system＇ s functions, to solve a class of high-dimensional optimization problem constrained. The key of algorithm is： the affinity of antibodies had relation to the antibody＇ s density and current population, the feasible antibodies were cloned and mutated, repaired the infeasible antibodies by means of the increasing sorting of price consistency of antibodies gene. In numerical experiments, selected two existing intelligent algorithms （ETGA , ISGA） to compare with the designed algorithm, tested high-dimensional 0/1 knapsack problems with different con- straints. The results indicate that the new algorithm can track rapidly the optimum, and also show the predominant exploitation and exploration capability of algorithm.
出处
《计算机应用研究》
CSCD
北大核心
2009年第8期2921-2923,2930,共4页
Application Research of Computers
基金
安顺学院青年一般项目基金资助(20080514)
关键词
高维0/1背包问题
约束优化
抗体修正
免疫算法
high-dimensional 0/1 knapsack problem
constrained optimization
antibody repair
immune algorithms
