摘要
利用人工免疫系统的学习、记忆、识别等功能,提出一种动态免疫优化算法(DIOA),用于解决一类高维动态约束优化问题。其中对可行抗体进行克隆突变操作,非可行抗体按价值密度使用贪婪算法进行修正,环境识别模块借助记忆细胞产生新的环境初始群,从而加快算法收敛速度。利用DIOA求解不同环境下的高维背包问题,结果表明,与同类算法相比,该算法能更快地跟踪最优值,收敛效果更好。
This paper proposes a Dynamic Immune Optimization Algorithm(DIOA) based on biological immune system learning,memory and recognition functions to solve a class of high-dimensional dynamic optimization problem with constraints.The feasible antibodies are cloned and mutated,the infeasible antibodies are repaired,by means of the profit-density of antibody,and the new environmental population is generated by using memory cells of similar environment,which accelerates the convergence of algorithm.The algorithm is applied in the high-dimensional knapsack problems are solved in different environments.Experimental results prove that,compared with traditional algorithms,DIOA can track the optimum rapidly and has stronger convergent capability.
出处
《计算机工程》
CAS
CSCD
北大核心
2011年第20期216-218,222,共4页
Computer Engineering
基金
贵州省自然科学基金资助项目(20090074)
关键词
动态环境
高维动态约束优化
背包问题
免疫优化
贪婪算法
dynamic environment
high-dimensional dynamic constraint optimization
knapsack problem
immune optimization
greedy algorithm