Delay system identification using global optimization methods

Delay system identification problem is studied in this paper. Converting to an optimization problem, a method using global optimization algorithms is proposed to identify concurrently all the parameters, including both continuous variables, such as the coefficients of the model, and discrete variable, time delay. Two global optimization methods, tabu search (TS) and simulated annealing (SA), are developed to deal with this problem. Simulation results show that both the TS and SA methods can identify the delay system successfully.