Designing feeder bus routes by using smart card data

Attempts have been made of over decades to design a more efficient and yet sustainable transportation system by using different mass transportation systems. A simple form of this, is the intermodal transportation system which can be described by feeder bus route design problem (FBRNDP). One of the fundamental components of transit planning is the estimation of origin-destination (O-D) matrix. Collecting such data, however, is extremely difficult and expensive using traditional paper-survey-based approaches which are rarely available. Therefore, a new data source was explored in the form of transit smart cards. This spatio-temporal data is most important for any meaningful transit smart card studies, and even so, if they are used alongside other data sources such as data base management system and geographic information systems, their potential can be harnessed. Multiple Traveling Salesman Problem (MTSP) is similar to FBRNDP in which there are many origins and one destination. The main purpose of this research is to develop a methodology for the planning of public feeder bus routes by using smart card data, transit network with existing bus stop coordinates and a distance optimization algorithm. In this study, important aspects of feeder bus route network design are highlighted, a method for the determining of potential feeder bus stop location developed, and a modified genetic algorithm was used to solve FBRNDP. The methodology was used to design a new feeder bus routes, and the best solution was selected via pareto efficiency and minimized total system cost.