基于RS-FNN的电梯群智能控制策略

在电梯群控制调度算法中，如何有效利用电梯群控制系统中的大量复杂、冗余或不完
备的原始数据以及群控制策略如何适应多变的电梯交通状况是提高电梯群控性能的关键问题。本文提出一种新型的粗糙集－模糊神经网络混合控制策略。针对电梯群控制系统中大量可参考属性信息，采用粗糙集理论进行属性约简，提炼出在不同交通流模式下对系统最重要的属性，再建立相应交通流模式下的模糊神经网络。仿真实验证明该群控制策略在派梯调度时，可实时预测和辨识电梯交通流模式，根据不同交通流模式采用不同的模糊神经网络模型进行派梯计算，尤其提高了电梯群控系统在混合交通流状况下的运行效率。
关键词:电梯群控系统；粗糙集；模糊神经网络
Abstract: Effective mining the huge elevator data in elevator group control system and improve the performance in variable traffic pattern is the key problem in elevator group control policy research. In this paper, a novel intelligent elevator group control strategy based on rough set theory and fuzzy neural network called RS-FNN strategy is proposed. As a strong data fusion method, rough set theory can extract the most important system attributes. This proposed control method firstly reduce the main attributes in elevator group control system with discernibility matrix method in rough set theory, and then build up a fuzzy neural network to control the elevator group. The simulation results show that this proposed novel elevator group control method can improve elevator system performance and adapt complicated traffic flow in variable elevator traffic flow conditions.
Keywords: Elevator group control system; rough set, fuzzy neural network