Neural network for modelling a physical system and method for building the neural network

摘要

The neural network has input, output and group layers. The number of groups used depends on the number of summation stems of the function with which the system is mathematically described. Groups of neurons are formed, the number of groups equaling the number of sub-functions of the equation describing the system and the sub-function coefficients are arranged in the form of untrainable input connections after the output neuron of a group. The neural network has an input layer (3) and an output layer (5) and a group layer (4) consisting of at least two groups, whereby the number of groups used depends on the number of summation stems of the function with which the system to be modeled is mathematically described. Groups of neurons are formed, whereby the number of groups is equal to the number of sub-functions of the functional equation describing the system to be modeled and the sub-function coefficients are arranged in the form of untrainable input connections after the output neuron of a group. AN Independent claim is also included for the following: a method of setting up a neural network, a method of training a neural network and an optimization method.