摘要
提出一种无圈有向图条件下的多目标最短路径进化算法。使用变长染色体对路径编码。进行染色体适应值分配时同时考虑支配关系及密度信息,保持了种群的多样性。有界精英保留策略保证了算法的优化性能。对算法的收敛性进行了证明。理论分析和实验表明,该算法可以在较短时间内获得多条多目标优化路径。
A multiobjective shortest path evolutionary algorithm is presented for acyclic directed graph. It employs variable -length chromosomes for encoding the path. Dominance relation and density information are incorporated to fitness assig ment for keeping diversity of individuals. Bounded elitism ensures optimal performance of the algorithm. The convergence behavior of algorithm is proved. Theoretical and experimental results show that the algorithm can obtain multiple multiobjective optimal paths within acceptable time.
出处
《系统工程》
CSCD
北大核心
2005年第9期123-126,共4页
Systems Engineering
基金
国家863高技术资助项目(2002AA783030)
关键词
多目标最短路径
多目标进化算法
支配
有界精英策略
Multiobjective Shortest Path
Multiobjective Evolutionary Algorithm
Dominance
Bounded Elitism