Distributed Multi-Agent System for Real-Time Traffic Optimization:A Comparative Analysis of Advanced Greedy Heuristics in UrbanCongestion Management

Abstract

Traffic congestion is a major problem in many regions, especially in large cities. Several factors contribute to this situation, such as inadequate road infrastructure and the high number of vehicles, particularly during peak hours: in the morning from 7 a.m. to 8 a.m., at noon from 12 p.m. to 1 p.m., and from 3 p.m. to 5 p.m.. Our main objective is to propose a practical and adaptable approach to managing congestion by using the capabilities of multi-agent systems to make quick and relevant decisions in real time. We aim to make traffic flow smoother, reduce pollution, and make driving more pleasant and safe for all road users. Our project proposes an agent-based simulation to manage road traffic. This approach involves the cooperation between different actors, such as drivers and traffic management systems, to optimize traffic flow and reduce congestion in cities. The overall waiting time of vehicles on the roads is then evaluated using different heuristics like FIFO (First In First Out) and two versions of the Greedy heuristic, one simple and the other advanced. In the end, a comparison between these three heuristics shows that the advanced Greedy heuristic better optimizes traffic management and reduces vehicle waiting times.