GENETIC-ALGORITHM-BASED MICRO-SIMULATION APPROACH FOR ESTIMATING TURNING PROPORTIONS AT SIGNALIZED INTERSECTIONS

This paper presents a method for the automatic estimation of turning proportions at signalized urban intersections using partial link count information. The method integrates a microscopic model to simulate the network traffic conditions with a genetic algorithm to minimize the difference between estimated and observed movement flows. The implementation into a real signalized arterial network showed that the method can reproduce observed flows with a reasonable accuracy and computing speed that allows its usage in real-time urban traffic control systems.