Multi-strategy Enhanced Particle Swarm Optimization Algorithm for Elevator Group Scheduling

Although particle swarm optimization has shown great potentials in solving the complex elevator group scheduling problem, it still suffers from the issue of local optimum. In order to improve the capability of finding the global optimum, a multi-strategy enhanced particle swarm optimization algorithm has been proposed for elevator group scheduling in this work. For the initialization of particle position, Tent map is used to generate a diverse position distribution for faster and more effective explorations in the entire solution space. Spiral flight strategy is then utilized to update the position and velocity of particles in a more flexible way with exploring more spaces. Once an optimum is obtained, a local search strategy is finally employed to search the nearby solution spaces to further avoid local optimum. Simulation results have demonstrated that the proposed algorithm can achieve a shorter passenger waiting time than traditional particle swarm optimization.