Scholarlink

<p>The inventory routing problem (IRP) poses a significant optimization challenge across various industries. This paper focuses on the uncapacitated IRP, by introducing fast constructive heuristics integrating insights from approximation algorithms, particularly rounding techniques in linear programming&#xa0;(LP). The proposed heuristics efficiently deliver effective solutions, providing advantages over methods such as branch-and-cut and metaheuristics. Methodologically, we emphasize scalability, subjecting our algorithms to rigorous stress tests with larger instances. Computational experiments, utilizing 420 instances, demonstrate the effectiveness and scalability of our heuristics, notably those tailored to specific problem variants, achieving an average gap of 2.2%. Our work underscores the effectiveness of leveraging approximation algorithms for the uncapacitated IRP, with future directions aimed at enhancing heuristics for broader real-world applicability, including the capacitated version of the IRP.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Fast constructive heuristics for the uncapacitated inventory routing problem

  • 期刊论文

  • 发表日期：

    2025年5月10日

查看全文
  • Miguel Ángel Marfurt Alarcón,
  • Lehilton Lelis Chaves Pedrosa,
  • Fábio Luiz Usberti

摘要

The inventory routing problem (IRP) poses a significant optimization challenge across various industries. This paper focuses on the uncapacitated IRP, by introducing fast constructive heuristics integrating insights from approximation algorithms, particularly rounding techniques in linear programming (LP). The proposed heuristics efficiently deliver effective solutions, providing advantages over methods such as branch-and-cut and metaheuristics. Methodologically, we emphasize scalability, subjecting our algorithms to rigorous stress tests with larger instances. Computational experiments, utilizing 420 instances, demonstrate the effectiveness and scalability of our heuristics, notably those tailored to specific problem variants, achieving an average gap of 2.2%. Our work underscores the effectiveness of leveraging approximation algorithms for the uncapacitated IRP, with future directions aimed at enhancing heuristics for broader real-world applicability, including the capacitated version of the IRP.