Forecasting model based on multidimensional moving pattern for a class of complex production process

For a class of complex production processes, in our previous work, the idea of moving pattern based modeling is proposed. But, in our previous work, the moving pattern is confined to one dimension and the one dimensional moving pattern-based modeling is studied. In this paper, moving pattern is extended to multidimensional case, and a multidimensional moving pattern based forecasting model is proposed. First, the algorithm for constructing multidimensional pattern moving space is proposed. Second, for characterizing pattern class variable quantitatively, the multidimensional interval autoregression model (MIAR) is defined. Third, the proposed MIAR is applied to modeling the movement of pattern class variable in pattern moving space. At last, experimental results are then presented that indicate the validity and applicability of the proposed model.