Frequency setting on public transportation networks for passenger and operator cost minimization using metaheuristic algorithm

Rapid growing cities in terms of population and surface emerges a lot of problems. Since rapid grow prevents their well-planned development. Especially, cities in developing countries have transportation-based problems such as long travel times, traffic congestion, accidents, air and noise pollution. In order to solve these problems, transit systems are one of the most effective methods. However, transit systems, particularly in countries having limited resources, should be used in optimal yield for creating sustainable cities. By only arranging frequencies, transit service can be provided more efficient usually without any investment cost. In this thesis, the frequency setting was proposed for the least cost of transit in terms of both operator and user. Bi level optimization model was used. In the upper level of bi-level structure, frequencies were determined using firefly metaheuristic algorithm while in the lower level, transit assignment was done on simulation program according to frequencies determined in the upper level. Transit demand was assumed to have been fixed. While being done transit assignment, the frequency based assignment was adopted. Also, it was aimed that the volume of each line does not exceed the capacity of each line in the given assignment period. Firefly algorithm and COM connection between Visum and Matlab was developed on the Mandl test network. Then, the model was implemented in a part of Izmir, Turkey. With the existing frequencies, a total cost of 145.201 TL was incurred, and 991 users were traveling over the capacity. In the end of optimization, the total cost was reduced to 138.029 TL with zero user over the capacity.