Maintenance Job

Maintenance job scheduling aims to optimize the timing and execution of maintenance tasks to minimize downtime and maximize efficiency across various systems, from datacenters to power generation units. Current research emphasizes the development of robust predictive models, often employing machine learning techniques like quantile regression and deep reinforcement learning, to accurately forecast job durations and account for uncertainties and safety constraints. These advancements are crucial for improving resource allocation, reducing operational costs, and enhancing the reliability of complex systems in diverse sectors. The integration of human expertise within these automated systems is also a growing area of focus, improving prediction accuracy and ensuring safe and effective maintenance strategies.

Papers

December 10, 2022

Acela: Predictable Datacenter-level Maintenance Job Scheduling
Yi Ding, Aijia Gao, Thibaud Ryden, Kaushik Mitra, Sukumar Kalmanje, Yanai Golany, Michael Carbin, Henry Hoffmann
Data Center Cloud Data Center Duration Prediction Maintenance Job

June 23, 2022

Human-in-the-Loop Large-Scale Predictive Maintenance of Workstations
Alexander Nikitin, Samuel Kaski
Domain Expert Predictive Maintenance Reachable Workspace Maintenance Job

December 20, 2021

Safe multi-agent deep reinforcement learning for joint bidding and maintenance scheduling of generation units
Pegah Rokhforoz, Olga Fink
Reinforcement Learning Deep Reinforcement Learning Electricity Market Safe Multi Agent Reinforcement Learning Maintenance Job

Maintenance Job

Papers

Acela: Predictable Datacenter-level Maintenance Job Scheduling

Human-in-the-Loop Large-Scale Predictive Maintenance of Workstations

Safe multi-agent deep reinforcement learning for joint bidding and maintenance scheduling of generation units