多目标遗传算法在电梯主动导轮系统集成优化设计中的应用  

Application of multi-objective genetic algorithm to integrated plant/controller optimization for active guide roller design

作　　者：傅武军[1] 朱昌明[1] 叶庆泰[1] 

机构地区：[1]上海交通大学机械与动力工程学院,上海200030

出　　处：《振动工程学报》2006年第2期227-233,共7页Journal of Vibration Engineering

摘　　要：介绍了一种基于多目标遗传算法的电梯主动导轮系统集成结构和控制器优化设计方法。在本方法中,电梯主动导轮结构参数和主动控制器参数同时作为优化设计变量,电梯系统性能和控制代价作为优化设计目标。这是一个复杂的多目标优化问题。由于控制器具有特定的结构,多目标遗传算法直接被用来求解该多目标集成优化问题。这样不仅可以避免权函数的选择,而且一次运行就可以求出所有的Pareto最优解。最后用一个实验室比例电梯模型作为例子来说明该设计过程。A new approach to design active guide roller based on Multi-objective Genetic Algorithm is proposed. In this / method, the structural parameters of the guide roller as well as the active controller are regarded as the design variables, and the system performance and the control cost are optimized. Because of the fact a structured controller is used, this simultaneous optimization problem is solved directly by the multi-objective genetic algorithms, by which a Pareto-optimal set is obtained within a single run, and furthermore the choosing of the weights, via which the relative importance of the optimization objectives is defined, is avoided. A scaled elevator in lab is used to illustrate the design process. And both the numerical and experimental results are presented to verify the efficiency of this method.

关 键 词：振动控制 多目标遗传算法 主动导轮系统 结构/控制器集成优化 

分 类 号：TB535[理学—物理] TH21[理学—声学]