Scale Forecast Method for Regional Highway Network Based on BPNN-MOP

The forecast of the scale of highway network is of great importance in the planning of regional highway network. This paper is to seek a hybrid method to improve the accuracy and reliability of scale-forecast and obtain the optimal hierarchical structure of highway network in Hangzhou in the Year 2015, 2020 and 2025. Firstly, drawbacks of traditional scale-forecast methods of highway network and advantages of the combinative forecast method which embraces BP neural network (BPNN) and Markov chain are illustrated. Then, a novel prediction method which is based on BPNN and Markov chain with the consideration of five elements, including GDP, total population, the number of civil motor vehicle ownership, passenger capacity and volume of freight traffic is proposed. After that, a multi-objective programming (MOP) model is established to obtain the optimum technical grade structure of highway network. Thirdly, with the Program for the 13th Five-Year Development Plan of Highway Transportation in Hangzhou, the scale of highway and its optimal hierarchical structure in the year of 2015, 2020 and 2025 is obtained. Finally, the results show that the accuracy and reliability of the forecast method are improved, and the model proves to be of both theoretical and practical significance. (C) 2017 The Authors. Published by Elsevier B.V.