Swarm Algorithm for Traffic

This article considers a problem of route simulation of autonomous vehicles in a collision-free traffic. A new swarm algorithm is proposed, based on a microscopic traffic model, which ensures the collision-free motion of agents. The change in the several optimality criteria is considered in the process of the algorithm operation, such as: average speed of agents, throughput, and the number of lane changes. The boundaries of the effective values of the algorithm hyperparameters are estimated. Free flow and an improvement in the optimization criteria are observed for certain density parameters and attraction-repulsion factors in the traffic.