An On-Demand Alternate Transportation Service System for Public Transportation using Real-Time Multi-agent Technology

For medium and large-scale urban rail services, any disruption of the operating schedule due to signal malfunction or accident involving injury or death can have wide ranging effects throughout the city. When such incidents occur, alternate means of transportation such as temporary bus services are arranged, but currently, operation of such buses is decided onsite. This research proposes a system that links agents in a multi-agent simulation to real-world transpor-tation service companies and passengers and is able to make predictions within several minutes of an incident, of the operating state and transportation conditions 1 to 3 hours later, to decide a plan for operating alternate transportation based on these predictions, and to present this plan to bus companies and passengers in an on-line fashion. The results of simulations using a prototype of this system indicated that it will be able to facilitate bus companies to provide alternate trans-portation, and passengers to use such transportation, within five minutes of an incident occurring in a public transportation service on the scale of a city of one million people.