A statistical learning based approach for parameter fine-tuning of metaheuristics

IF 0.7 4区数学 Q4 OPERATIONS RESEARCH & MANAGEMENT SCIENCE

Sort-Statistics and Operations Research Transactions Pub Date : 2016-06-17 DOI:10.2436/20.8080.02.41

Laura Calvet, A. Juan, C. Serrat, Jana Ries

引用次数: 28

Abstract

Metaheuristics are approximation methods used to solve combinatorial optimization problems. Their performance usually depends on a set of parameters that need to be adjusted. The selectionof appropriate parameter values causes a loss of efficiency, as it requires time, and advanced analytical and problem-specific skills. This paper provides an overview of the principal approaches to tackle the Parameter Setting Problem, focusing on the statistical procedures employed so far by the scientific community. In addition, a novel methodology is proposed, which is tested using an already existing algorithm for solving the Multi-Depot Vehicle Routing Problem.

查看原文本刊更多论文

基于统计学习的元启发式参数微调方法

元启发式是用于解决组合优化问题的近似方法。它们的性能通常取决于一组需要调整的参数。选择合适的参数值会导致效率的降低，因为它需要时间、高级分析和特定问题的技能。本文概述了解决参数设置问题的主要方法，重点介绍了科学界迄今为止采用的统计程序。此外，提出了一种新的方法，并利用已有的算法对该方法进行了测试，以解决多车场车辆路线问题。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

求助全文

约1分钟内获得全文求助全文

来源期刊

Sort-Statistics and Operations Research Transactions 管理科学-统计学与概率论

CiteScore

3.10

自引率

0.00%

发文量

审稿时长

>12 weeks

期刊介绍： SORT (Statistics and Operations Research Transactions) —formerly Qüestiió— is an international journal launched in 2003. It is published twice-yearly, in English, by the Statistical Institute of Catalonia (Idescat). The journal is co-edited by the Universitat Politècnica de Catalunya, Universitat de Barcelona, Universitat Autonòma de Barcelona, Universitat de Girona, Universitat Pompeu Fabra i Universitat de Lleida, with the co-operation of the Spanish Section of the International Biometric Society and the Catalan Statistical Society. SORT promotes the publication of original articles of a methodological or applied nature or motivated by an applied problem in statistics, operations research, official statistics or biometrics as well as book reviews. We encourage authors to include an example of a real data set in their manuscripts.