|
The traveling salesman problem (Traveling salesman problem) is a classic combinatorial optimization problems, is also an NP-complete problem, very popular around the real life application. Applications such as in the printed board manufacturing, ultra-large-scale integrated chip manufacturing, intelligent control of the industry, the application of robotics research and control and other related areas. Take into account the wide range of uses, scientists and researchers have been trying to seek a, which reflects the high-quality results, but also fast convergence to the optimal or approximate calculation algorithm. The calculation algorithm of the traditional areas of the branch and bound method, the support tree redouble law, greedy algorithm and partial closing cable law, in recent years, in some places the bionic algorithm, including ant algorithms, simulated annealing, genetic algorithms and neural networks such as some of the more advanced algorithm, compared with traditional algorithm, the degree of convergence of these algorithms to some extent, have been greatly improved, the quality of the results is also improved. The genetic algorithm (Genetic Algorithm, GA) calculation algorithm is developed through simulation of biological natural selection and evolution principle based on the probability of the global search. With other methods do not have a lot of advantages in many methods to solve the traveling salesman problem, genetic algorithm, genetic algorithm can obtain the optimal solution for small and medium-scale traveling salesman problem, can get the approximate optimal solution for large-scale traveling salesman problem. In solving the traveling salesman problem, the ant algorithm as a heuristic algorithm exhibited excellent performance. The the outer hormone concentrations preferably path can to strengthen, the next selected path in accordance with the the outer path hormone concentration is selected, thereby increasing number of ants will select the best path through ants secretion of hormones, better path will be covered by more ectohormone all ants eventually focus on a good path. The basic principle of the whole algorithm is ant-based pheromone positive feedback principle is the key point of the algorithm lies. The genetic algorithm has a fast overall, search capabilities, but the lack of full use of the feedback system is running, the result will have no use tedious iterative, which greatly reduces the efficiency of solving. , Ant colony algorithm is ultimately achieve the purpose of convergence to the optimal path through the cumulative update ectohormone. Therefore, the genetic algorithm with distributed, parallel, global convergence ability. Early relative lack of pheromones, resulting in the slow process of algorithm. Their defects in order to overcome the genetic algorithm and ant algorithm to achieve mutually complementary, complementary advantages, this paper take the route of the first use of genetic algorithms random search quickly, a whole generation of convergence on the issues related to the initial pheromone distribution Next, take full advantage of the highlight a bit of the ant colony: a positive feedback mechanism, parallel and solving high efficiency characteristics, take this mixed calculation algorithm inspired are the better quality in terms of time efficiency, or in solving type algorithm. In-depth study of the genetic algorithm and ant algorithm, analysis, and ultimately this paper the combination of the genetic algorithm and ant colony algorithm fusion TSP problem solving, related to the specific work carried out as follows: 1. Inspection, read a large number of professional literature, analysis of the traveling salesman problem arising from the background, domestic, international research progress of this study's purpose, significance, while the line of research presented in this article, and for work. Expounded the concept of the traveling salesman problem, the mathematical model of the classification and the traveling salesman problem, Intro and analyzing several classic traveling salesman problem solving algorithm, genetic algorithm and ant colony algorithm and its characteristics, the basic principles of research status quo and its improved algorithm by more than the contents of the study, an overview of the basic rules of the existing hybrid genetic algorithm to analyze the advantages and disadvantages of hybrid genetic algorithm through analysis study of several hybrid genetic algorithm to solve the traveling salesman problem, proposed kinds of new hybrid algorithm, and a detailed description and analysis of the algorithm design and coding scheme, simulation experiments obtained simulation results verify the hybrid algorithm feasibility, scientific, and ultimately concluded: new optimize the efficiency and quality of the algorithm are better than the basic ant colony algorithm and genetic algorithm.
|