Method and arrangement as well as a computer program comprising program code - means -, and computer program product for the determination of a future system state of a dynamic system

摘要

The invention relates to a neuronal arrangement, i.e. a recurrent neuronal network structure suitable for predicting the future state of a dynamic system and to a mapping which describes the dynamic system performance and suitable for predicting the dynamic system future state. For this purpose, the system state variables are exposed at the first stage to a spatial transformation which transforms space state variables into a transformation space. The state variable transformed into the transformation space are exposed to a time transformation which transforms the transformed state variables in the transformation space from the first moment to a second future moment. The state variables transformed at the second future moment are exposed to a space transformation which retransforms the state variables transformed at the second future moment of the transformation space into the space state. The state variables of the state space transformed into the transformation space and the state variables of the space transformation transformed into the actual space state comprise the influence quantities of the dynamic system and at least one final quantity of the dynamic system influenced by the influence quantities. The transformation parameters for carrying out the space transformations are fixed whereas the transformation parameters for carrying out the time transformations are variable.