A competitive analysis approach for route choice with uncertain travel times and blocked nodes

Abstract

This paper investigates the route choice problem on a network with uncertain travel times along the arcs and possible blocked nodes. The route choice decision is made with the simultaneous consideration of these two kinds of nondeterministic factors. In order to deal with these nondeterministic factors, online algorithms are introduced to develop adaptive route choice strategies responding to the successively observed blocked nodes; and accordingly, an uncertain competitive analysis framework, taking expected competitive ratio as the metric, is proposed to evaluate the performances of the online algorithms associated with uncertain travel times. It is proved that the competitive analysis for this route choice problem can be handled in the framework of traditional online shortest path problem via its equivalent counterpart on the corresponding deterministic network. Furthermore, two typical online strategies, i.e., reset strategy and greedy strategy, for route choice decision-making and the resulted performance difference are analyzed in detail following from the proposed framework.