Design of IDSS for Traffic Microsimulation Model Calibration

Abstract

Traffic microsimulation is a widely used tool in the field of computer-aided traffic engineering. In procedure of building traffic microsimulation model, the step of model calibration is difficult. Although several factors of calibration has been discussed by scholars, the calibration step has been handled by means of the experience of individual traffic engineer up to now. This unsystemic approach way makes the traffic microsimulation modeling unreliable and inefficient. In this paper, we take the process of microsimulation model calibration as a decision-making problem, and design an IDSS (Intelligence Decision Support System), named VisController, to aid this decision-making process. In VisController, we integrate several preceding researches about traffic microsimulation model calibration into a system. Two main computer technologies are combined in VisController to make this system intelligentized: neural network and genetic algorithm. Neural network are used in pattern recognition of a micro simulation model; genetic algorithm is used to identify the optimal parametric values. In addition, a demo, which is based on Vissim, a widely used traffic microsimulation software, has been presented.