Moving pattern-based forecasting model of a class of complex dynamical systems

Considering the existence of uncertainty of complex dynamical systems, in contrast to traditional modeling approach to characterizing dynamical systems in Euclidean space, the dynamics of complex industrial processes is characterized by the move of operating condition patterns in pattern moving space. First, the operating condition patterns of complex dynamical systems are partitioned into C pattern classes constructing pattern moving space, and then pattern class variable characterizing the movement of operating condition patterns in pattern moving space is defined. Each pattern class characterized (quantified) by an interval-valued number can be considered as the “calibration” in pattern moving space. For modeling the move of pattern class variable in pattern moving space, interval autoregression model (IAR) is defined and applied to modeling the movement of pattern class variable in pattern moving space. Finally, Experimental results are then presented that indicate the validity and applicability of the proposed approach.