Author(s)

Haoyu Wang, Chengfei Yang and Hui Zhou

Corresponding Author

Haoyu Wang

ABSTRACT

Bus dispatching is a crucial issue in bus operation, and a suitable dispatching method can significantly save operating costs. This paper proposes a passenger flow prediction and division method based on BP neural network and Fisher optimal dissection method for the dynamic dispatch of buses. Based on this method, the bus dispatcher can increase or decrease the number of buses before peak traffic according to the results of prediction and division to achieve economic efficiency and other purposes. This paper uses the passenger flow data of a certain line in Guangzhou for a month to train and test the network and obtain good results.

KEYWORDS

Traffic Flow Prediction, Bus Dispatching, Bp neural network, Fisher Clustering Method