Incorporating User Acceptance Probabilities in Optimizing Profit of Carsharing Systems

Bekli, Seyma and Boyacı, Burak and Zografos, K. G. (2022) Incorporating User Acceptance Probabilities in Optimizing Profit of Carsharing Systems. In: 15th International Conference on Advanced Systems in Public Transport, 2022-11-06 - 2022-11-10, Israel. (Unpublished)

Abstract

Station-based one–way electric carsharing (SBOWECS) systems provide a convenient, flexible, and environmentally friendly alternative to accommodate demand for short (in terms of time and space) trips. A key determinant of the level of service, and profitability of SBOWECS systems is the availability of adequately charged vehicles to accommodate the spatial and temporal characteristics of the demand. Vehicle relocation is used to ensure that vehicles are available at the right time at the right place to serve customer trip requests. Therefore, the cost of the personnel involved in vehicle relocation operation constitutes a significant part of the operating cost of SBOWECS. One way of dealing with the cost of the relocation operations is to reduce the need for relocating vehicles by providing incentives to potential users in order to better align the spatial and temporal characteristics of demand (trip requests) and supply of vehicles. Studies with alternative offers in the literature assume that users accept all the offers made by the system regardless of their specific spatial, temporal, and price characteristics. However, in reality, the users may decline alternative offers if the time, space, and price characteristics of the offers do not align well with their preferences. With this study, we aim to develop a decision framework that would consider the probability of users rejecting an alternative offer. In the decision framework, we assume that the users are placing trip requests to the system, and the offers are made by the operator without the knowledge of future requests. Each user has a predefined acceptance probability which is a function of the "distance" between the request and the offer. As soon as a request is received by the operator, the proposed decision-making framework evaluates the request and makes a counteroffer to the user that maximizes the expected marginal profit. In addition to the SBOWECS system parameters, we use the acceptance probability distribution of the users, previously accepted requests, and the historical demand to calculate the best offer.