The Study And The Application Of ACO In The Control System

ACO is put forward by M.Dorigo in the early1990s.It is a new type of heuristic algorithmwhich is a real simulation of ant colony foraging behavior in the nature. It is a heuristic searchalgorithm for combinatorial optimization problems after the neural network geneticalgorithm,simulated annealing algorithm, genetic algorithm, the ACO algorithm,etc.In the pasttime, Ant colony algorithm has been widely used in combinatorial optimization, data mining,function optimization and network routing, system identification, and other fields.This paper is around the principle of ant colony algorithm and theory, the application of antcolony algorithm in the control system has carried on the thorough system’s research. In thispaper, the main research results include:1).Put forward a method of the satisfactory PID parameters optimization based on ant colonyalgorithm. When controlled variable is a constant value, Traditional PID parameter optimizationuses Z-N optimization method, etc.When controlled variable is Interval variational, PIDparameters optimization commonly uses the robust algorithm.But he robust algorithm is strongtheoretical,it is difficult to grasp.So Put forward a method of the satisfactory PID parametersoptimization based on ant colony algorithm in this paper.The simulation results show thefeasibility and effectiveness of the algorithm.2). Put forward a method of solving the nonlinear equations based on ant colonyalgorithm.The solution of the nonlinear equations generally uses the numerical methods. Theconvergent speed of this algorithm is faster.But for the solution of some strong nonlinearequations, numerical method is easy to result in failure and low efficiency.So put forward amethod of solving the nonlinear equations based on ant colony algorithm in this paper. Thesimulation results show the effectiveness of the algorithm. The solving speed is higher than thetraditional numerical methods’s.3).Put forward a method of the Wiener model identification based on ant colony algorithm.Due to the diversity of the static nonlinear function and the uncertainty of intermediatesignal,it’sdifficult to do the Wiener model identification. When the nonlinear module of theWiener model is complicated, the parameters of the model could be as high as dozens of,even more. The use of traditional optimization algorithm is easily into local optimum.It’s almostimpossible to get the correct solution.So put forward a method of the Wiener model identificationbased on ant colony algorithm. The simulation results show the feasibility and effectiveness ofthe algorithm.