Transfer Optimization in a Transit Network

Abstract

Transfer optimization attempts to minimize the overall inconvenience to passengers who must transfer between lines in a transit network. Bus trips are scheduled to depart from their terminal so as to minimize some objective function measuring that inconvenience. In this paper, the transit network is assumed to be given, and the scheduled headway is treated as fixed on each line. We denote by ti the departure time of the first bus on line i. {ti} are termed “offset times,” and constitute the decision variables of our model. To take into account stochastic travel times of buses, our treatment of transfer optimization employs a simulation procedure in combination with an optimization model. That model turns out to be a relaxation of the Quadratic Assignment Problem. It can incorporate a wide range of objective functions (measures of overall passenger disutility) and a variety of policies for holding buses at a transfer point. In the case where buses are not held at all, we show, for a number of different objective functions and transit networks, the negative consequences of optimizing transfers with a deterministic bus-travel-times assumption, if these travel times are in fact random variables. Suggestions are then made for future research.