Integrated Swarm Intelligence Framework for Dynamic Traffic Optimization in Delhi: A Three-Layer PSO-Fuzzy-MAS Approach

Abstract—This study presents a hybrid, three-layer framework for intelligent traffic optimization in Delhi, combining Particle Swarm Optimization (PSO), Fuzzy Logic Controllers (FLC), and Multi-Agent Systems (MAS). It tackles challenges posed by real-time traffic fluctuations, dense vehicle networks, and limited infrastructure scalability. The system uses PSO for global and MAS for self-organized traffic agent collaboration. SUMO simulations with 628 vehicles show a 32.1% reduction in average travel time and a 28.3% drop in fuel use. Field testing at ITO Junction validated 41% peak-hour congestion reduction. The system exhibits real-time adaptability, decentralized decision- making, and significant environmental impact reduction. Index Terms—Swarm Intelligence, PSO, Fuzzy Logic, MAS, Traffic Optimization, Delhi