Surrogate-assisted cooperative signal optimization for large-scale traffic networks

Reasonable setting of traffic signals can be very helpful in alleviating congestion urban networks. Meta-heuristic optimization algorithms have proved themselves to able find high-quality signal timing plans. However, they generally suffer from performance deterioration when solving large-scale problems due the huge search space and limited computational budget. Directing against this issue, study proposes a surrogate-assisted cooperative (SCSO) method. Different existing methods that directly deal with entire network, SCSO first decomposes it into set tractable sub-networks, then achieves by cooperatively optimizing these sub-networks optimizer. The decomposition operation significantly narrows whole optimizer greatly lowers burden reducing number expensive simulations. By taking Newman fast algorithm, radial basis function modified estimation distribution algorithm as decomposer, surrogate model optimizer, respectively, develops concrete algorithm. To evaluate its effectiveness efficiency, network involving crossroads T-junctions is generated based on real network. Comparison several meta-heuristic specially designed for demonstrates superiority average delay time vehicles.