Approaches to control acyclic traffic lights in an exemplary urban road network

Abstract

This paper presents a comparison of five algorithms used to control acyclic traffic lights at intersections of roads in an urban road network. The following algorithms are selected: Most Cars characterized by low computational complexity, the author's algorithm called In-and-Outbound Lane Control, which is an efficient modification of the Most Cars, Local Hill-Climbing algorithm (LHC), the reinforcement learning RL 1 Bucket 2.0 algorithm and the neural network GenNeural algorithm.