t-CSA: A fast and flexible CSA implementation

Abstract

Interactive journey planning on public transport networks is a challenge. We humans cope with the journey planning challenge by intuitively searching timetables for patterns. Connection scan algorithm (CSA) is a promising algorithm to automate interactive journey planning by searching timetables for best possible journey plan. We propose t-CSA algorithm as an extension of CSA. Our contribution is two fold; Our first contribution is the inclusion of direct connection table (DCTable) in CSA to accelerate the discovery of frequent direct connections. Our second contribution is the inclusion of all known multi-modal connections between nearby stations in journey planning by using other means table (OMTable). OMTable permits customization of journey plan as per the transport facilities available to a user. Our algorithm has an average response time of 0.7ms when run on timetables of Indian public transport networks which includes both Indian railways and bus networks. We find that our algorithm makes interactive journey planning applications feasible.