Identification Of Van Der Pol-Duffing Oscillators Based On Artificial Bee Colony Algorithm

System control and synchronization are main components of control engineering. They are under spotlight of many areas in engineering application. As a prerequisite of system control and synchronization, system identification is of great importance in both theory and application. Since1967, the International Federation of Automatic Control (IFAC) has held a number of system identification and parameter estimation thematic sessions. It has been made a lot of achievement in system identification. System identification not only has a wide range of applications in the aerospace, industrial production, but also plays a very important role in the field of economic management, bio-medicine, meteorology, environmental engineering and social systems, etc.As a new kind of swarm intelligence optimization algorithm, artificial bee colony (ABC) algorithm was proposed in2005. It has been studied only for several years. The algorithm is described for solving multidimensional and multimodal optimization problems and its simulating results are comparable with differential evolution algorithm and particle swarm optimization algorithm. Because of the advantages of few parameters, fast evolution speed, simple calculation and easy implementation, the algorithm has highly attracted the attention of researchers since its birth. But until now, little research has been done to identify systems through this algorithm.This paper is concerned with further analysis of ABC algorithm. Based on the analysis, we will modify the basic ABC algorithm, propose a proper mathematical model, design a common method for solving system identification problems, then identify unknown parameters of the nonlinear vibration system "Van der Pol-Duffing (VDPD)" oscillator with improved ABC algorithm.In chapter1, we introduce the research background and meaning of system identification firstly. Then the methods of identifying system and research status at home and abroad are given. Thirdly, the main aspects, purpose of this research and structure arrangement are introduced.In chapter2, we roughly described the concept and meaning of intelligent optimization algorithm. And then, several algorithms of them are listed. At last, evolutionary algorithms and swarm intelligence algorithms are highlighted through a detailed introduction of GA, DE, ACO, PSO based on optimization principle, algorithm processes, current research and applications of the algorithms. In chapter3, we discourse ABC algorithm in detail. First, the ABC algorithm generation and development process are given. Second, optimization principle of this algorithm is illustrated. Finally, the global convergence of this algorithm is simply introduced.In chapter4, we propose ABC algorithm with space contraction strategy. To improve exploration ability of the later generation, we define a new search space centering current local optimum after a certain iterations of evolution. After that, we initialize a new population in the new search space.In chapter5, we identify unknown parameters of VDPD oscillators with ABCSC algorithm. The identification process is divided into two stages:VDPD oscillators with and without noise. Based on quantitative and qualitative analysis of the identification results, feasibility and effectiveness of improved ABC algorithm in resolving such problems are indicated.In chapter6, the work carried out in this paper is summarized. And the work to be done in the future is discussed.