Title: Multi-level Statistical Models for Asset Management and Logistics
Speaker: Maharshi Dhada, Assistant Professor, IIT Indore
Date & time: Oct 26, 2026 Monday, 15:30-16:30
Venue: IEOR 211 (Seminar Room), 2nd floor, IEOR building
Abstract: This seminar presents multi-level statistical models for inference in machine fleets where individual data are sparse. A fleet-wide model obtained by pooling all assets is often biased by heterogeneity in operating conditions, whereas independent asset-specific models exhibit high variance when the data are limited, such as for newly commissioned assets. Multi-level models address this bias-variance trade-off by drawing asset-specific parameters from higher-level distributions shared across clusters of similarly behaving assets. Maharshi will discuss hierarchical Weibull models for times to failure, the propagation of parameter uncertainty into remaining useful life and maintenance policies, the incorporation of expert knowledge through the higher-level structure, and adaptive inference for sequential updating. This will be supported with case studies on commercial vehicle and turbofan fleets. He will also present the extension of this framework to demand forecasting for new locations in logistics networks, evaluated by the operational cost of the resulting decisions, and conclude with his research and teaching plans.
Speaker Bio: Maharshi is an Assistant Professor in the Department of Mechanical Engineering at IIT Indore, where he heads the Collaborative Industrial Systems Group. He has held the prestigious SF Express Research Fellowship in Logistics at the University of Cambridge. Maharshi holds a BTech in Mechanical Engineering from IIT Indore and a PhD in Engineering from the University of Cambridge. He studies decision-making under uncertainty in industrial operations, with applications in logistics and engineering asset management where data are distributed across individual components of the systems. Maharshi has collaborated with many leading institutions and organisations such as MIT, UCL, Siemens, Boeing, Scania CV, BT, Glovo, and SF Express.