为求解多目标0-1背包问题,基于竞争决策算法原理和多目标优化问题的特性,提出了一种求解多目标0-1背包问题的元胞竞争决策算法。将元胞自动机演化规则引入竞争决策算法,给出了算法的具体描述,并使用Delphi7.0实现了算法的具体步骤。为了提高多目标非劣解(Pareto解)的分布性和多样性,利用全局经验作为指导,在最稀疏的Pareto解附近进行邻域搜索。经过大量数据测试和验证,该算法具有真实的Pareto前沿逼近效果,是一种多目标优化问题的有效方法。
In order to solve the multi-objective 0-1 knapsack problem,based on the mechanism of CDA( competitive decision algorithm,CDA for short) and the feature of multi-objective optimization problem,this paper presented a cellular CDA for multi-objective 0-1 knapsack problem. It introduced the rule of cellular automata into CDA and gave its mathematical description. Coded the algorithm in Delphi 7. 0,by which tested series of typical instances. In order to improve the diversity and distribution of Pareto optimal set it carry out neighborhood search among the most sparse Pareto optimalset. The simulation results show that the algorithm can efficiently reach the true Pareto frontier and the effectiveness of the algorithm.