期刊
ANNALS OF OPERATIONS RESEARCH
卷 253, 期 2, 页码 825-843出版社
SPRINGER
DOI: 10.1007/s10479-016-2177-5
关键词
Tabu search; Uncertainty; Scenario; Routing
资金
- DOMinant II Project
- Research Council of Norway
- Iceland, Liechtenstein
- Norway through the EEA Financial Mechanism [026-ABELIM-2013]
- Government of Spain [TIN2012-32482]
Heuristics for stochastic and dynamic vehicle routing problems are often kept relatively simple, in part due to the high computational burden resulting from having to consider stochastic information in some form. In this work, three existing heuristics are extended by three different local search variations: a first improvement descent using stochastic information, a tabu search using stochastic information when updating the incumbent solution, and a tabu search using stochastic information when selecting moves based on a list of moves determined through a proxy evaluation. In particular, the three local search variations are designed to utilize stochastic information in the form of sampled scenarios. The results indicate that adding local search using stochastic information to the existing heuristics can further reduce operating costs for shipping companies by 0.5-2 %. While the existing heuristics could produce structurally different solutions even when using similar stochastic information in the search, the appended local search methods seem able to make the final solutions more similar in structure.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据