多目标随机规划的交互遗传算法  

Interactive Genetic Algorithm for Multiobjective Stochastic Programming

作　　者：胡毓达[1] 杨雷[1] 等 

机构地区：[1]上海交通大学应用数学系，上海200030

出　　处：《上海交通大学学报》2001年第11期1733-1736,共4页Journal of Shanghai Jiaotong University

基　　金：国家自然科学基金资助项目 ( 70 0 710 2 6)

摘　　要：利用遗传算法在处理过程中不依赖问题的种类 ,并具有较强鲁棒性等特点 ,提出了一种基于交互式的求解多目标随机规划的遗传算法 .算法的思想是 ,结合小生境技巧和构造 Pareto选优过滤器的手段 ,通过与决策者的反复交互对话 ,最后得到使决策者满意的问题的 Pareto有效解集 .The increasing complexity in decision making process has brought new hard solved problems involving diversity of objectives and various random factors. The generic algorithm (GA) is referred to as an efficient parallel and evolutionary search technique. Because of its independence of problem types in actual models and its better robustness in the iterative process, GA plays an important role in successfully handling complicated multiobjective problems. In this paper, a newly developed stochastic multiobjective genetic algorithm was introduced on basis of interactive approach. Integrated the niche technique with the construction of Pareto set filter, through continuous interaction with decision maker, a new family of Pareto efficient solution which satisfies the decision makers could be obtained.

关 键 词：多目标随机规划 交互规划算法 遗传算法 随机模拟 Pareto有效解集 运筹学 

分 类 号：O221[理学—运筹学与控制论] O242.23[理学—数学]