Optimizing a demand-responsive feeder system for low-demand areas

Abstract

Efficient public transport systems are essential for the dynamism of medium and large cities. With the growth of urban areas around large centers, new neighborhoods emerge in the suburbs and pressure the public transport system. A limited solution is to extend the system with feeder lines. However, in the last decades, demand-responsive systems emerged, complementing traditional systems and providing efficient transportation services for users and operators. We introduce a demand-responsive public bus system that considers shortcuts and detours and modifies the departure times of a feeder system, based on the current demand. This system contains a recognizable backbone, but it is not fixed. The operation is optimized, using a memetic algorithm, for each operation period in order to reduce passenger travel time. The optimized system in our experiments reduces passenger travel time with 18 to 38% compared to a traditional system operating the same network with fixed routes and timetables.