Using variance neighbourhood search to optimize the bus waiting allocation problem in a multi-floor bus station

Abstract

This research deals with a multi-floor bus station, which provides services for a large number of passengers. The bus station has a limited platform capacity and there is no temporary parking lot for buses. When a large number of buses move into one floor of the station, buses cannot move smoothly and may not even be able to move at all. When a floor is full, buses waiting outside cannot enter, and buses inside cannot move out. It is fortunate that the station is designed as a multi-floor structure. When a bus is scheduled to move onto a floor, which has no more space for parking, it can move to another floor temporarily to wait. This research proposes the use of integer-programming to optimize the assignment of temporary waiting floor for all incoming buses in order to minimize the maximum delay. A Variance Neighbourhood Search (VNS) is proposed to solve the problem. The results show that when temporary waiting on another floor is permitted, the total time delay can be reduced by up to 47.41%.

Keywords

• bus station
• bus waiting allocation
• variance neighbourhood search