Open AccessFleet Maintenance Strategy Planning with Time Windows Integrated with Multi-Agent and Wolf Pack Reinforcement Learning
Author(s) -
Ma Xinrui,
Li Haixu,
Zheng Zhang,
Bo Chen
Publication year - 2021
Publication title -
iop conference series. materials science and engineering
Language(s) - English
Resource type - Journals
eISSN - 1757-899X
pISSN - 1757-8981
DOI - 10.1088/1757-899x/1043/4/042036
Subject(s) - reinforcement learning , plan (archaeology) , computer science , operations research , function (biology) , foraging , motion planning , route planning , fleet management , artificial intelligence , engineering , transport engineering , ecology , archaeology , evolutionary biology , biology , robot , history , telecommunications
Selective maintenance is a widely used strategy to identify and perform the maintenance actions necessary for fleet mission success. Aiming at the problem of maintenance strategy planning with time windows (MSPTW) which is common in short-term operation plan, a fleet maintenance strategy planning approach based on multi-agent and reinforcement learning is studied in this paper. Based on the four kinds of foraging behaviors including migration, summon and attack in traditional Wolf Pack Algorithm (WPA), the intelligent behavior is redefined, and a new wolf pack algorithm for solving the MSPTW is designed. In order to seek the best path planning, a mathematical model with the aim of minimizing the total cost (fixed cost, transportation cost, waiting cost and penalty cost) is constructed utilizing the fitness and penalty function.