The University of Toronto Operations Research Group (UTORG) is hosting a lunch talk by Mahsa Hosseini. The talk is entitled “Efficient Repositioning in Car Sharing Networks”. Lunch and coffee will be provided. Hope to see you there!
Who: Mahsa Hosseini, Ph.D. candidate, University of Toronto
When: Wednesday, March 18th @ 12:10pm – 1:00pm
Bio-sketch: Mahsa Hosseini is currently a Ph.D. candidate at the Rotman School of Management, University of Toronto. She is supervised by Prof. Joseph Milner and Prof. Gonzalo Romero. She holds an MBA from the Department of Management and Economics at Sharif University of Technology. Mahsa’s research interests include stochastic control, stochastic modeling, and theory of Markov decision process.
Abstract: We propose a modelling framework to study the dynamic problem of repositioning vehicles using steady-state behaviour for a network with centralized control and uncertain, unbalanced demand. To provide interpretable strategic level reposition policies, we use the structure of the underlying transition matrix to estimate the relocation gradient as a function of the current number of vehicles in each location. In particular, we project the full-dimensional problem of finding optimal open-loop relocation policy onto the lower-dimensional space of reposition decisions only. Then we segment this space and estimate the local change to the steady-state and interpret the meaning of these gradients via absorbing Markov chain concepts. These gradients then provide a state-dependent heuristic for the dynamic vehicle repositioning problem.