Research On Intelligence Of Urban Bus Dispatching Methods

The urban citizens' travel demand of China is increasingly expanding,resulting from the rapid economic development,urbanization,rising population and motor vehicles in the past few decades,and the problems of traffic congestion and environmental pollution have been aggaravating.Public transportation has received attention gradually with the intensive study in the field of transportation.Meanwhile,the development of public transportation is regard as an important part in solving problems of city development and building sustainable city.Bus dispatching system is the core of transit system,which means it plays a key role in improving the level of service of public transportation system.This paper proposes a new idea of intelligence of bus dispatching methods,so as to promote the efficiency of the bus dispatching system.Firstly,the investigation of characteristics of the transit dispatching system is based on reading literatures of this field widely,and pros and cons of the existing dispatching modes and methods are concluded.Meanwhile,the demands of dispatching system and existing problems are also analyzed.The summary of basic data which include passenger demands and buses operating status is according to the demand analysis,and the intelligent dispatching model is set up based on the data.Secondly,the division of time intervals before modeling is considered while most existing study lack consideration of the specific impact of time distributiong of buses operating status and passenger demands.In addition,the dispatching model with and without time intervals division are compared,and the results show that time intervals division is effective in decreasing dispatching work loads and raising efficiency.Finally,the core content of this investigation is the remission of problems in prediction of bus operating bias and response of dispatching by the intelligent methods.Accordingly,the dynamic dispatching model based on DNN(Deep Neutral Network)is built after the division of time intervals and static scheduling,in order to realize the fast recognition of bus operating bias and fast execution of dispatching strategies.The model is tested according to the data obtained from the field survey.The results show that the proposed intelligent dynamic dispatching model can decrease the passenger waiting time and instability of bus headways effectively with no addition of delay.