Scholarlink

The Green Two-Level Vehicle Routing Problem (G2E-VRP) is an extension of the VRP and is an NP-hard problem. Traditional approaches use independent single-task paradigms without considering task similarity. To address this, an Adaptive Evolutionary Multi-Task (AEMT-PMKS) algorithm based on the Pyramid Matching Kernel Strategy (PMKS) is proposed. First, the two-level transportation network of the G2E-VRP is clustered to construct a multi-task framework. A memetic algorithm is used to handle each task, and the Pyramid Matching Kernel Strategy is introduced to enable resource sharing among similar tasks. During the evolutionary process, short paths are detected, and the relationship between users and satellites is dynamically adjusted to obtain the global optimal solution. The widely used 2E-VRP benchmark instances are adopted, the effectiveness of the AEMT-PMKS algorithm is verified.

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

Adaptive Evolutionary Multitasking with Pyramid Matching Kernel Strategy for Solving the Green Two-Echelon Vehicle Routing Problem

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Nannan Zuo,
  • Rong Hu,
  • Qingxia Shang,
  • Yuxiao Huang,
  • Bin Qian

摘要

The Green Two-Level Vehicle Routing Problem (G2E-VRP) is an extension of the VRP and is an NP-hard problem. Traditional approaches use independent single-task paradigms without considering task similarity. To address this, an Adaptive Evolutionary Multi-Task (AEMT-PMKS) algorithm based on the Pyramid Matching Kernel Strategy (PMKS) is proposed. First, the two-level transportation network of the G2E-VRP is clustered to construct a multi-task framework. A memetic algorithm is used to handle each task, and the Pyramid Matching Kernel Strategy is introduced to enable resource sharing among similar tasks. During the evolutionary process, short paths are detected, and the relationship between users and satellites is dynamically adjusted to obtain the global optimal solution. The widely used 2E-VRP benchmark instances are adopted, the effectiveness of the AEMT-PMKS algorithm is verified.