Modeling the Allocation of Urban Bus Fleet Using ANNs (Case study: Mashhad)

Using Inter-city bus system with regard to high flexibility and low cost for the users, especially in large cities and in developing countries, is inevitable. So transportation system improvement is necessary due to its range of activity. One of the ways in improving this system is optimizing the allocation of buses to active lines that compared with present status, carries more number of passengers and also its transporting cost will be reduced. In this study, a new approach for optimizing the allocation of buses to the existing bus network lines is used. In this method, a model is employed which can determine the number of required buses for line by considering various parameters. For this purpose, artificial neural networks were used. Artificial neural networks due to the ability of learning with examples can be used for non-linear modeling in which exact solution is hard to achieve. After completing the training process, network performance was investigated and when satisfying the constraint conditions, it could simulate new cases. In this investigation, the optimized network (NN 11-13-1) with maximum R-value (R=0.996) and minimum squared error (MSE=0.553) was used. Finally, the model obtained on the bus network in Mashhad was tested. The results obtained using neural networks in comparison with regression shows its high level of accuracy which is a valuable approach.