Input-output structure decomposition for multivariable control system by genetic algorithm

In this paper, a input-output weighting matrix is proposed as the coupling measure between the input variables to the output variables. From the structure of the input-output weighting matrix, we can reduce the inter-channel coupling by permutation and blocking. This paper shows how to use this information to identify the strongly and weakly coupled sub-systems by setting the partition of the input-output weighting matrix. A Genetic Algorithm for this application is proposed to determine the Lest permutation and the number of the partition, so that the inter-channel coupling will be deduced after rearranging the input/output variables.