Identification of Continuous-Time Systems from Sampled Data via Genetic Algorithm

摘要

This paper proposes a system identification method of a continuous-time system from sampled input and output data using a genetic algorithm(GA). A gene-coding based on a pole-zero parameterization is adopted for a plant to be identified. It is shown that a GA is able to directly identify poles and zeros. It is obvious that an ordering difficulty arise from the pole-zero parameterization. To avoid the difficulty, we introduce a certain ordering procedure before crossover. Finally, an illustrative numerical example is given.