Date: Wednesday, October 9th, 2019, 5:15PM-6:15PM
Title: Operating Battery Swap Stations: A Scheduling, Allocation, and Inventory Replenishment Approach
Abstract: Drones, Electric Vehicles (EVs), and other battery-dependent systems promise to transform many industries from drone delivery to critical infrastructure monitoring. However, the reality of these transformations can be hindered by short battery operation times and life-spans. Furthermore, the act of enabling a battery to operate in the short term through recharging is harmful for long-term operation due to battery degradation. Short-term recharging is necessary to enable batteries to complete tasks and satisfy demand. However, recharging causes batteries to degrade with deteriorating long-term performance which can only be reinstated with costly battery replacement. Thus, we must consider inventory replenishment decisions that balance short-term recharging decisions with long-term battery replacement. We consider operating a battery swap station in a central location that enables drones and EVs to swap depleted batteries for full ones in a matter of seconds. To operate the swap station, we must determine how many batteries are charged, discharged using battery to grid, and replaced over time. Additionally, we must allocate and schedule batteries for the completion of tasks or satisfaction of demand. Thus, we propose the new class of scheduling, allocation, and inventory replenishment problems (SAIRPs). For SAIRPs, we introduce new deterministic mixed integer programming formulations and stochastic Markov Decision Processes. We present the results about the complexity and structure of optimal policies. To solve these complex models, we outline new heuristic and approximate dynamic programming methods. We show the results based on operating a drone delivery system with a single swap station used to deliver blood to hospitals in Rwanda.
Bio:Sarah Nurre Pinkley received her PhD from the Industrial and Systems Engineering Department at Rensselaer Polytechnic Institute. Prior to this, she earned a Masters in Industrial and Management Engineering and a Bachelor’s in Mathematics, both from Rensselaer Polytechnic Institute. Prior to joining the University of Arkansas in August 2015, she was an Assistant Professor in the Operational Sciences Department at Air Force Institute of Technology. Her research interests include using network optimization, scheduling, and optimization algorithms for restoring interdependent infrastructure systems, operating Electric Vehicle and drone battery swap stations, and understanding last-mile delivery. She has taught undergraduate courses in Operations Research, Master’s courses in Optimization Theory, Multicriteria Optimization, Stochastic Modeling and Analysis, and Heuristics, and PhD courses in Scheduling Theory and Linear Programming. Sarah is actively involved in many INFORMS organizations including Women in Operations Research and the Management Sciences (WORMS), Subdivisions Council, and the Professional Recognition Committee. At the University of Arkansas, she serves as the faculty advisor for the Institute of Industrial and Systems Engineers student chapter.