Many existing methods for transit assignment and path finding do not support schedule coordination in network search or optimization processes. Other methods do not provide adequate performance that meets the demands for Internet trip-planning applications. Two schedule-based path-finding algorithms (forward search and backward search) are presented for the transit network, in which schedule coordination is an inherent feature. The forward-search algorithm finds the optimal path from an origin to a destination with a planned departure time, and the backward-search algorithm finds the optimal path for an expected arrival time at the destination. In addition, a non-schedule-based minimal-transfer-path algorithm for the transit network is also developed, which is capable of accommodating more path optimization criteria for sophisticated path finding. Facilitated by a uniquely designed network structure based on a geographic information system and an object-oriented data model, these algorithms demonstrated good performance in path finding on a dynamic transit network. These algorithms have been implemented in an Internet transit trip-planning system in the city of Waukesha, Wisconsin.
Samenvatting