A MIP Model to Optimize Real Time Maintenance and Relocation Operations in One-way Carsharing Systems

Abstract

The daily management of a one way carsharing system is focused on relocation and maintenance operations. On one hand, the freedom of allowing users to return vehicles wherever they want leads to vehicle imbalance problems and on the other hand, a system where the same vehicle is used by different persons, has the need for daily maintenance operations. Demand is a key element that influences both relocation and maintenance needs, and the system has to be prepared to adapt and react to it in an optimized way. This paper describes the optimization model integrated in a tool developed to manage relocation and maintenance operations of a one-way carsharing system in real time. The optimization model developed is a mixed integer linear programing model, instant-based and designed to work using a rolling horizon. The division of the operation time into planning horizons allow the updating of the system status data, which provides a closer connection to the system reality. The model considers that crew elements use mainly the vehicles of the system to move inside the operating area, and orders are remotely transmitted through a wireless communication platform to mobile devices. Three types of activities are attributed to crew elements: waiting, maintaining the vehicles and moving. The mathematical model discriminates each crew element and is able to decide the best schedule for each one. It has flexibility to select between crew trip joining or crew trip splitting, in order to reduce movement cost or relocate a higher number of vehicles. Simulation tests were performed to assess the computer processing time for different problem sizes.