Scholarlink

This chapter focuses on optimizing condition-based maintenance (CBM) for deteriorating systems. It reviews relevant scientific literature on CBM modeling and optimization up until 2024. The chapter primarily examines papers that utilize Markov Decision Processes (MDPs) for optimization, categorizing them based on their approach to analytically clarify the properties of an optimal policy and to computationally obtain the optimal policy. The chapter begins with an explanation of the essential elements of maintenance planning: the deterioration process, maintenance actions, and associated costs. The review covers both single-unit and multi-unit systems, noting the increased complexity and dependencies in multi-unit systems. Additionally, the chapter discusses the joint optimization of maintenance with other decisions, such as spare parts ordering and mission abort strategies, emphasizing the potential of deep reinforcement learning for large-scale CBM problems. The chapter concludes with future research directions, advocating for algorithms that leverage structural properties to enhance CBM policies.

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

Two Approaches to Condition-Based Maintenance Optimization for Deteriorating Systems

  • 章节

  • 发表日期：

    2025年

查看全文
  • Lu Jin,
  • Mizuki Kasuya

摘要

This chapter focuses on optimizing condition-based maintenance (CBM) for deteriorating systems. It reviews relevant scientific literature on CBM modeling and optimization up until 2024. The chapter primarily examines papers that utilize Markov Decision Processes (MDPs) for optimization, categorizing them based on their approach to analytically clarify the properties of an optimal policy and to computationally obtain the optimal policy. The chapter begins with an explanation of the essential elements of maintenance planning: the deterioration process, maintenance actions, and associated costs. The review covers both single-unit and multi-unit systems, noting the increased complexity and dependencies in multi-unit systems. Additionally, the chapter discusses the joint optimization of maintenance with other decisions, such as spare parts ordering and mission abort strategies, emphasizing the potential of deep reinforcement learning for large-scale CBM problems. The chapter concludes with future research directions, advocating for algorithms that leverage structural properties to enhance CBM policies.