An Optimization Method Combined Genetic Algorithm with Neural Network

Based on genetic arithmetic and neural network theory, aggregate gradation of asphalt stabilized base course mixtures was optimized. In the course of optimization, the target function was asphalt mixtures fatigue properties, and the decisive parameter were the weight passed through 9.5 mm sieve and ore powder dose. Compared to Super pave aggregate gradation, the optimized one fixed in with Super pave-gradation prescript totally. Through fatigue experiment, the optimized asphalt mixtures fatigue properties was the longest. It is shown that the optimization method based on genetic arithmetic and neural network can be used to optimize asphalt mixtures aggregate gradation. This method is available to optimize involved target function that cannot be expressed by the decisive parameter apparently.