Multi-objective scheduling algorithm for distributed elevator group control system
分布式电梯群控系统多目标调度算法

The distributed elevator group control system is based on the multi-agent system(MAS). The elevators and the group-controller are mapped respectively to different agents with specific functions; the hall calls are dispatched to elevators through negotiations among agents, thus the computation of scheduling algorithm is decentralized to different agents. Based on the market-like model, the distributed group-control algorithm includes mainly the quotation-generation algorithm of C-agents and the elevator dispatching algorithm of HC-agent. A distributed multi-objective group-control algorithm(DMO) is designed to optimize the average waiting time(AWT), the average riding time(ART), and the long waiting percentage(LPT). Simulation results show that the proposed MAS-based distributed elevator group control system is feasible, and the DMO algorithm has high performance in optimizing AWT, ART, and LPT.