基于鱼群算法的电梯群控调度算法

Elevator group control dispatch algorithm based on artificial fish-swarm algorithm

为了提高电梯群控系统(EGCS)运行效率,针对国内外电梯群控调度算法存在的问题,提出了一种将电梯轿厢内人数作为调度因素,综合考虑乘客平均等待时间、平均乘梯时间、电梯能耗的智能调度算法。将人工鱼群算法理论与实际的电梯群控调度问题相结合,通过对离散鱼群算法参数的优化、步长的改进,使其适用于电梯群控调度。该算法首先初始化一个鱼群,使鱼群中每条人工鱼当前的状态代表一种可行的调度方案,在经轿箱内人数调度因素缩小后的解空间范围内,让鱼群通过其觅食行为、聚群行为和追尾行为进行全局寻优。最后,把该算法与最小等待时间调度算法进行了比较。研究结果表明,乘客的平均候梯时间、平均乘梯时间、长候梯率较最小等待时间调度算法有明显降低。

In order to improve the operating efficiency of elevator group control system （EGCS） and aiming at the deficiency in current EGCS scheduling algorithm at home and abroad, an intelligent scheduling algorithm was proposed, the number of people inside the elevator was taken as the scheduling factor, the average passenger waiting time, average riding time and energy consumption were considered. The artifi- cial fish swarm algorithm （AFSA） theory was applied into practical elevator group scheduling problem, the discrete fish swarm algorithm pa- rameters were optimized and the step length was improved in order that the scheduling algorithm was appropriate for the elevator group sched- uling. The algorithm first initialized a shoal of fish, in which each fish＇s present state represented a viable scheduling scheme. Then, it opti- mized generally the foraging behavior, poly group behavior and the rear-end behavior in space scale after reduction of number scheduling fac- tors in the lift box. Finally, this algorithm was compared with the minimum waiting time scheduling algorithm. The research results indicate that the average waiting time, average riding time and frequency of long-waiting of passengers are significantly reduced.