A novel bus scheduling model based on passenger flow and bus travel time prediction using the improved cuckoo search algorithm

Abstract

Public transport is vital to people's daily travel and work, and bus scheduling plays a significant role in public transportation. A complete bus scheduling model can reduce the waiting time of passengers at bus stops and take into account the interests of bus companies. Currently, many bus scheduling models only predict the total passenger flow of a bus route but fail to predict the passenger flow of each stop of the bus route and the bus travel time between stops. Targeting the above challenges, first, we use LSTM, CNN-LSTM, and ConvLSTM to predict the passenger flow and travel time. Second, based on the predicted results, a bus scheduling model that minimizes the passengers' waiting time and the bus company's cost is established. Furthermore, a cuckoo algorithm with adaptive step size, nonlinear inertia weight logarithmic decreasing, and randomly adjusted discovery probability (AS-DWCS) is proposed and employed to solve the optimal bus scheduling plan. Finally, our experimental results show that the improved cuckoo algorithm converges faster and can seek the optimal solution faster. The bus scheduling model can consider the interests of bus companies, minimize passenger waiting time, and dynamically schedule the bus simultaneously.