【摘要】 针对一类未知动力学方程的单输入系统，采用插值的粗粒化 Ｂｏｘ Ｍ ｏｄｅｌ进行学习控制．与传统的Ｂｏｘ Ｍｏｄｅｌ及基于随机元胞自动机的控制模式比较，本模型设计思想自然，算法简单．计算量小（常常小几个数量级），训练时间短，对三个较有代表性的实例的仿真结果也是令人满意的更多还原

【Abstract】 In this paper, learning control is applied to a single-input system with unknown dynamical equations, using; coarse- grained Box Model with interpola- tion. As compared with the traditional Box Model or with Stochastic Cellular Automata, this model is more reasonal in idea, much simpler in algorithm, less in computation amount and less training time. Three illustrative simulation examples are given and show satisfactory results.更多还原