Design of parameter estimator using ant colony system

In this paper, a parameter estimator using adaptive ant colony system (adaptive ACS) algorithm is proposed for improving the precise parameter of the controller which is difficult to obtain. Ant family, a cooperative agent algorithm, has characteristics of positive feedback, distributed computation, and the use of a constructive greedy heuristic. Hence, the searching pattern of ACS algorithm is a discrete type such that it can not be suitably employed in the parameter estimation of some uncertain system. In order to overcome these problems, the adaptive ACS algorithm is proposed. And, nonlinear functions are adopted in each layer of searching patterns of adaptive ACS algorithm for rapid searching of the global optimal solution. By tuning the parameter of nonlinear function, the parameter estimation of uncertain system can be quickly obtained by the adaptive ACS estimator in the initial state. Moreover, the proposed estimator provides real-time estimation value of parameter of the controller. Finally, the effectiveness of the proposed adaptive ACS algorithm has been verified by simulations.