Operational Effects of Service Level Variations for the Dial-a-Ride Problem

Abstract

The dial-a-ride problem consists of designing a number of minimum-cost vehicle routes in a system of demand-dependent, collective people transportation. Quality is ensured by taking into account service level requirements. However, little research has been performed regarding the effect of simultaneous service level variations on operational costs incurred by dial-a-ride service providers. By investigating 78 different quality scenarios, being a combination of a maximum deviation from the user's preference time and a maximum relative exceedance of the direct ride time, this study presents a clear tradeoff between quality and costs, which enables providers to make an informed decision with regard to potential changes in the service level they offer. Results are found to be dependent on the size of the provider. Additionally, effects of traffic congestion and heterogeneous user types are investigated. The financial impact of the results is illustrated by means of an example. The entire study is performed on well-known benchmark data from literature and employs the deterministic annealing metaheuristic of Braekers et al. (2014) to compute operational costs.