Intelligent Elevator Control Using Decentralized Multi-Agent Systems

Abstract

A multi-agent system (MAS) framework optimizes elevator operations in high-rise buildings with dynamic traffic patterns and fluctuating passenger demands. With the proposed MAS approach, which utilizes fully decentralized decision-making and dynamic task allocation, scalability, responsiveness, and energy efficiency are significantly improved. Comparing MAS to traditional centralized systems, simulation results show that it reduces average waiting times by up to 25% while maintaining high performance. MAS eliminates bottlenecks and central points of failure, allowing real-time adaptation to traffic changes and passenger behavior. This approach provides significant advantages over existing elevator control strategies, positioning it as an intelligent building management solution of high effectiveness.