Ant Colony Optimization And Its Application To Blind Equalization

Ant colony optimization, a new meta-heuristic which has the advantage of robustness, global optimization, universality and distributed computation, has been used to solve complicated problems in combinatorial optimization. As the deepening of basic research and expansion in application field, ant colony optimization has showed its superiority to solving many combinatorial optimization problems. In the field of mobile communications, blind equalization algorithm based on wireless channel has undergone great changes. Some approaches such as CMA without common zeros, TXK, Subspace Algorithm (SSA), and Linear Prediction Algorithm (LPA) have developed to directly Blind detection algorithm which transmits signals with a prior information of finite alphabet and could be used in channels with common zeros. Just in light of current research and characteristics of the algorithm, ant colony optimization algorithm is applied directly to blind signal detection in order to further demonstrate its applicability and validity. Simulation shows that ant colony algorithm can successfully solve the problem of blind detection. The new algorithm, proposed in this thesis, can better avoid premature stagnation algorithm and optimize its performance.The thesis is organized as fellows. In chapter I, the main work of this thesis is described. Chapter II summarizes the research of domestic and international ant colony algorithm, introduces its basic algorithm in detail, and analyzes its convergence. Chapter III illustrates the blind SIMO system identification model, basic theory and evaluation indicators of blind equalization. Chapter IV presents the applications of the basic ant colony optimization algorithm in blind equalization. That is, transfer the problem of blind equalization into quadratic programming problem of finite alphabet and solve this problem with ant colony algorithm. Three improved algorithms are firstly proposed in chapter V, called blind detection algorithms based on TACO, CACO, and DACO. The advantage of these algorithms has been proved by research and simulation in terms of the bit error rate, complexity, and convergence. Chapter VI is a summary and outlook.