Research On The Theory Of Artificial Bee Colony Algorithm And Its Application On Communication

Optimization problems exist widely in many fields such as engineering technology, management science, computer science, scientific research and so on. At present, the common optimization algorithm for solving the optimization problems can be divided into classical optimization algorithm, local search algorithm and greedy algorithm, intelligent optimization algorithm, hybrid optimization algorithm and so on. Swarm intelligence optimization is an important branch of intelligence optimization. Swarm intelligence optimization is inspired from the collective behavior of social insects and is realized by the communication and cooperation between individuals. Compare with classical optimization algorithm, intelligent optimization algorithm has a lot of advantages, such as easily operating, fast convergence, good global searching ability, strong robustness and so on.Artificial bee colony (ABC) algorithm is a swarm intelligence optimization algorithm based on the particular intelligent behavior of honeybee swarms. Because of the advantages of its simple calculation, less control parameters, fast convergence and strong robustness, it has received more and more attention by scholars. The application of ABC algorithm and the improved methods are studied in the thesis.Firstly, the biological model, basic principle, implementation process and procedure of ABC algorithm are detailedly described, including characteristics and current research at home and abroad. Based on the advantages of ABC algorithm, a method for solving TSP problem by ABC algorithm is proposed. TSP problem is an NP-hard problem. For detail, the position of employed bee is mapped to a route; the food source is corresponding to the length of the route. Simulation results of14cities show that ABC algorithm is reliable and effective for solving combinatorial optimization problem.Secondly, Niche Artificial Bee Colony Algorithm (NABC) is proposed. Simulation results on two benchmark functions indicate that it is an effective algorithm comparing to Niche Genetic Algorithm for solving Multimodal Optimization problems.Thirdly, in the later iteration, algorithm has low convergent speed and population diversity seriously decreases, leading to be prematurely convergent and be trapped into local optimum. So in this thesis, Gaussian mutation and chaos disturbance are introduced into ABC to overcome the shortcomings above. An improved ABC algorithm has been proposed. In the local search, Gaussian mutation is carried out for improving the searching efficiency and precision. Considered the characteristics of ergodicity and randomness of chaotic variables, chaos disturbance is introduced into basic ABC, which is helpful for bees to jump out of local optimum and increases the population diversity. Simulation results on four benchmark functions indicate that the searching properties including searching efficiency and precision of improved ABC algorithm are obviously better than that of basic ABC algorithm, for solving global optimization problem.Finally, a method using ABC algorithm to optimize spectrum allocation for efficiency and fairness is presented. The experimental results show that not only can the proposed algorithm drastically improve network performance, but also the proposed algorithm provides benefits comparable to the Genetic Algorithm in terms of convergence speed, success rate and solution accuracy, meanwhile drastically reducing computation complexity.