Based Gm Model Ic Card Sub-time Passenger Flow Forecast

With the rapid development of intelligent transport system in China, traffic intelligent card is widely used thanks to its reliability, convenience, rapidity and other advantages. The crucial part for advanced urban public transport management is to seize change pattern of passenger flow. Therefore, it is of great practical significance and social value to calculate passenger flow information through analysis of IC data.This paper discusses the methods of data collection and the theoretical knowledge of passenger flow forecasting and gray model, researches and validates the methods of passenger flow forecasting based on GM model. This paper compares four methods of bus passenger data collection and points out the practicality and superiority of IC data acquisition.A method using residual GM model (GM1,1) for prediction is considered to forecast passenger volume in same period of a day accurately and scientifically. GM1,1 model based on IC data collected by single bus route is obtained. By determining coefficients of white equation, response time sequence of bus passenger volume is determined. Using response time sequence of gray differential equation to obtain restored value, development sequence of interval based passenger volume of a single bus route is established. Taking a variety of factors which influence bus passenger volume into consideration, development sequence is revised. Finally, the revised predicted value sequence is obtained.Numerical studies are made using IC data of bus passenger volume of JINAN city, to receive bus passenger volume prediction. It is shown that prediction model in this paper has high prediction accuracy; as a result, provides guidance to decision-makers of bus management.