Large-scale network signal control optimization method based on trust region Bayes

The invention belongs to the technical field of intelligent traffic signal control, and particularly relates to a large-scale network signal control optimization method based on trust region Bayes. The method mainly comprises three parts: model preparation, trust region Bayesian optimization and iteration termination. Wherein the model preparation part is mainly used for constructing a large-scale signal control optimization model, obtaining a priori training set and initializing a trust region; according to the trust region Bayesian optimization part, prior data set optimization is obtained based on the model preparation part, a next batch of sampling points (multiple sets of signal timing schemes) are obtained, a result is input into a microscopic traffic simulation model to operate, multiple sets of traffic evaluation target function values are obtained, and then whether optimization of iteration is terminated or not is judged through the iteration termination part. 本发明属于智能交通信号控制的技术领域，具体为一种基于信赖域贝叶斯的大规模网络信号