Finding alternative paths in city bus networks

Abstract

Finding alternative paths to follow is essential to travelers in dense city bus networks. A simple method is to transform the city bus networks into directed graph and then apply a standard K-shortest-path algorithm to the graph to find alternative paths. However, as the constructed graph is often massive and complex, classical K-shortest-path algorithms usually search for alternative paths that share a large proportion of links with the shortest path. For traveling purpose, there is not much difference among these alternative paths. We therefore develop a constrained K-shortest-path algorithm, in which a dissimilarity measurement is used as an additional constraint, to tackle this issue. This dissimilarity measurement is proposed to measure the diversity among alternative paths. Some experiments have been conducted based on the public transport network of Ho Chi Minh City (HCMC). The empirical results show that there exist only a small set of alternative paths that are indeed distinguishable from each other.