Planning limits failures while learning cuts costs in maintenance
Planning or Learning: Reliability and Cost in Multi-Asset Maintenance
Artificial IntelligenceMachine Learning
Summary
Keeping machines running reliably without spending too much is a hard balancing act. The authors compare two ways to schedule maintenance: one plans with strict rules to avoid any failures, while the other learns from experience and sometimes accepts small failures to save money. Planning makes sure machines never fail but costs don’t change much even if failures are very expensive. Learning methods can save costs when failures aren’t a big deal but still have some failures even if they are costly. The authors suggest which method suits different needs and share a setup for others to test these approaches.
What this means in practice
- •For maintenance planners: Create strict maintenance schedules that prevent all failures and keep reliability very high over short deployment periods.
- •For industrial operators: Run learned policies that tolerate occasional failures but reduce overall maintenance costs when some downtime is acceptable.
Authors
Xian Yeow Lee, Chandrasekar Venkatraman, Ahmed Farahat
Abstract
Industrial maintenance systems involve multiple interacting assets and shared resources, making it challenging to balance reliability and operational cost using a single decision framework. While recent work has focused on reinforcement learning (RL) for maintenance scheduling, direct comparisons with planning approaches under identical settings remain limited. In this work, we empirically compare planning and RL for multi-asset bearing maintenance using run-to-failure data. We examine how these methods behave when balancing preventive maintenance against tolerable failures across a range of failure penalty scenarios. We observed a consistent behavioral difference driven by objective formulation. Planning enforces reliability as a hard constraint and produces zero-failure policies whose total cost is largely insensitive to the magnitude of failure penalties. RL agents optimize expected cost and often trade off preventive maintenance against occasional failures as penalties vary, resulting in lower costs under low-penalty regimes but persistent non-zero failures even when penalties are high. We also investigate lightweight constraint mechanisms, including reward shaping and action masking, to encourage RL's reliability. From a practical perspective, planning may be more suitable when strict reliability is required and deployment horizons are short, whereas RL may provide cost-efficient policies when limited failures are acceptable and long-run operational efficiency is prioritized. Overall, this study clarifies the trade-offs between reliability and cost in multi-asset maintenance and suggests that planning and RL are complementary approaches. Beyond these findings, the controlled benchmark protocol itself that unifies environment, cost model, and evaluation across paradigms, offers a reusable template for comparing decision-making approaches in other maintenance settings.