This paper studies two methods of optimizing parameters of automatic detection for traffic incidents by using global search and implicit parallel of genetic algorithm .On the account of measured data of a real freeway. Firstly, the variables of California algorithm are optimized by using the Adaptive Genetic algorithm. Secondly, genetic algorithm based on real code is used to train the weigh value of artificial neural network for incident detection. Last, the testification of algorithm validation was undertaken. Research results indicate that genetic algorithm can optimize parameters of traffic incident detection well, and improve the efficiency of incident detection greatly. |