Application & Research Of Chaos Optimization Technique On Fuzzy Control System

Phenomena of chaos exists everywhere, it opens out the same behavior of nonlinear science, that is oneness between rigid regularity and randomicity, and oneness between regularity and out-of-order character. Ergodic, randomicity and regular properties are the characteristic of chaos, which means it can track any state in itself scope without repetition, according to its regularity. With the deeper research of chaos theory, as the core contents of nonlinear science research, the chaos application has become one of important issues and forefront project, and has been paid great attention in recent year, many desired research results have been achieved in the study of chaotic theory and its application, chaos optimization is one of important fields in chaotic scientific research, too.With the advance of industrial process automation and the deeper of optimal method in field of control, not only the high production efficiency, but also the high quality and low costs are required, a lot of practical control projects are summed up into optimal problems of the parameter of controller. And with increase of the complex industrial processes, it is very difficult to get the precise mathematical model because many practical problems of control project have the nonlinear, strong-constraint, random and large scale characters etc, and immanence mechanism is very complex. Some conventional optimal methods have disadvantages of tendency to become trapped in local minimum easily and slow convergence if initial values are not suitable, so the optimal effect can't reach request of system.Chaos optimization is an effective method that makes use of chaotic variables for optimal search, using the features of ergodicity and randomness of chaotic motion does the search process, it can continually search for the optimum solution, and overcome the local minimum problem. Compared to the stochastic search, chaos optimization seems to realize an efficient search for a variety of optimization problems, it can fast search global optimal solution.In this dissertation, the chaos theory and optimal method are roundly studied, the new chaos optimal method is proposed, and the innovatory researches are chaos optimization methods of fuzzy control system. In the end, an intelligent control system of model vehicle based on chaos optimization is designed.The main contributions of this dissertation are as follows.In chapter 1, the optimal methods and optimal theory and optimal problems about control system are particular studied. The characteristics of optimization method base on differential and intelligent are discussed, the status quos of chaos optimization in China and abroad are introduced.In chapter 2, the chaotic phylogeny, some basic definitions and conceptions are brief introduced; the Lyapunov exponent character of Logistic map is analyzed. We can find from figure of Logistic map, when parameter μ is changed, the track of system occur doubling period bifurcation and come into chaos finally. The theory of chaos dynamic system is analyzed, set out from definition and theorem, evolvement rule of Logistic map is discussed, some definitions and theorems of chaos theory are presented and proved, the dissertation provides basic theories of chaos optimization by quantificational analyzing the ergodicity and statistic properties of chaos series generated by Logistic map.The studies of chaos arithmetic of function optimization are discussed in chapter 3, in order to avoid blind searching of chaos optimization in searching space, an improving imitative space chaos optimization algorithm is proposed. The algorithm counts better value for every searching and sets a sign A in the chaos searching, when the numbers of better value searched is equal to A, the searching space is dynamic reduced according scale, and the above course is repeated in the lesser scale till global optimal value is found. The simulation results show that algorithm is simple and local searching ability is better, and the efficiency is higher than that of imitative scale chaos optimization.It is difficult to tune parameters of controller of fuzzy control system, and control rules and membership functions of FC are hard to obtain optimization too, therefore, the control results show strong overshoot andoscillation often. Chaos optimization methods of fuzzy control system are detailed researched in chapter 4. The fuzzy controller is constructed based on chaos algorithm off-line optimize the parameters and conjugate gradient descend method in-line tune the parameters. For inverted pendulum system, its four-dimensional output is directly decomposed into input of a pair of two-dimension fuzzy controller, the system is double-loop controlled, and the inner-loop regulates the angle of pendulum, while the outer-loop is for positioning the cart. The parameters optimization of controllers adopt chaos method, simulation result show that the system is steady. And than, the fuzzy neural network controller is designed based on chaos global optimize the structure and parameters, gradient descend method partial search the parameters. Simulations prove that optimization programs are simple and system is high control precision.In chapter 5, a control system of self-propelled sea-bed service vehicle is designed, which adopt closed loop of double loops in series. Fuzzy neural network is controller of outside loop, and fuzzy controller is inside loop. The structure and parameters of fuzzy neural network and parameters of fuzzy controller use off-line chaos optimization arithmetic, and then, the parameters optimized are connected to controllers. In the end, the conjugate gradient descend method is used to in-line tuned the parameters of fuzzy neural network in order to improve self-adaptive ability, thereby control effect is good.In chapter 6, the research results are concluded and further development is pointed out.