Choice-driven dial-a-ride problem for demand responsive mobility service.

Urban mobility services face the challenge of planning their operations efficiently while complying with user preferences. In this paper, we introduce a new mathematical model called a choice-driven dial-a-ride problem (CD-DARP) which is a generalization of the dynamic DARP where passenger behavior is integrated in the operational planning using choice models and assortment optimization. We look at two types of mobility services, private and shared. Our problem extends the dynamic DARP by (i) changing its objective function to profit maximization, where both cost and revenue are variables, and (ii) incorporating assortment optimization with routing decisions in a dynamic setting. We propose a pricing scheme based on a choice model designed to offer service alternatives at the time a customer makes a request. We introduce a tailored algorithm to efficiently solve the dynamic CD-DARP. Computational results indicate that our proposed approach outperforms dynamic DARP in terms of reducing routing costs and improving the number of customers served. • A new framework is introduced for DDARP incorporating customer choice. • The routing cost is reduced while satisfying customers' requests. • CD-DARP provides more flexibility for the operator to efficiently manage the fleet. • We evaluate the proposed model on synthetic as well as New York taxi data. • Flexible time windows helps improving the service level. [ABSTRACT FROM AUTHOR]