A bi-level approach to frequency optimisation of public transport systems

In this study, we concentrate on one of the three major areas of the whole transit network design problem that being the frequency setting. The main problem addressed here is the minimisation of losses borne by a public state transport corporation in India. This is achieved by optimally allocating resources to shared routes already in existence. The bus frequency setting problem is solved using a bi-level methodology. In the first level minimum required fleet size for the routes are found out by considering the routes individually. In the second level with the guarantee of load feasibility, allocation of frequencies is done for the fleet size found in the first level. This is done by also taking into consideration the achievement of minimisation of operational cost of the routes under consideration as added objectives. The relation between the operations cost and the maintained frequencies are found out using artificial neural network (ANN), and the second level is solved using a multi objective genetic algorithm (GA). This methodology has been used in optimising two shared routes and the results are thus presented.