摘要
双层CARP优化问题不仅要解决微观路径优化问题,还要解决宏观配置优化问题,最大程度地降低整体系统的固定成本和运行成本.提出了一种求解双层CARP优化问题的知识型蚁群算法:构建了一个动态参数决策模型,并采用该模型为每次迭代动态地选择一组合适的参数;基于弧段聚类知识和弧段顺序知识来构建可行解;采用2-Opt方法对每次迭代中的最优解进行局部优化,实验结果表明知识型蚁群算法在优化性能方面优于其他几种方法.
The double layer capacitated arc routing problem considers a high-level configuration prob- lem and a low-level routing problem, and its objective is minimize fixed costs and running costs of the whole system. A Knowledge-based Ant Colony Optimization （KACO） was proposed to the Double-layer Capacitated Arc Routing Problems. The exploitation of heuristic information, dynamic parameter adjust- ment and local optimization characterized the KACO. The dynamic parameter adjustment decreased the sensitivity of parameters to final experimental results. The feasible solution was constructed with the guid- ance of arc cluster knowledge and arc priority knowledge. Local optimization based on two-Opt heuristic largely improved the performance of KACO. In order to validate the performance of KACO, 87 benchmark problems were solved by KACO and some heuristic methods. Experimental results suggest that KACO outperforms these methods.
出处
《系统工程理论与实践》
EI
CSSCI
CSCD
北大核心
2012年第11期2540-2549,共10页
Systems Engineering-Theory & Practice
基金
国家自然科学基金重点项目(71031007)
国家自然科学基金(70971131
71101150
70801062)
高等学校博士学科点专项科研基金(20104307120019)
关键词
弧段顺序
弧段聚类
动态参数调整
宏观配置优化
微观路径优化
蚁群算法
arc priority
arc cluster
dynamic parameter adjustment
high-level configuration optimization
low-level routing optimization
ant colony optimization
