Intelligent Algorithms Based On Complex Network For Solving The Traveling Salesman Problem

TSP problem is an NP-hard problem which studied extensively in the field ofcombinatorial optimization, is also the centralized summarized and simplified form ofmany complex optimization problems in real life. It is widely used in a number of areas,such as industry, agriculture, national defense, construction, commercial, transportation andso on. Therefore, an effective solution for the TSP will drive the development of the societymore than the industry, with a wide range of practical value. In addition, due to the TSPproblem is one of the typical problems in the field of combinatorial optimization. Therefore,it is often used to evaluate the pros and cons of various heuristic optimization algorithms.Thus, the study of the TSP also has important theoretical value.With the development of computers, the researchers had given a lot of intelligentalgorithms for solving TSP problem. For example, genetic algorithm, ant colony algorithm,particle swarm algorithms and so on, these algorithms are basically evolved by some lawsof nature. The difference between tranditional work and the proposed algorithm is that thelatter uses some of the characteristics of complex networks for optimizing combinatorialproblems. The complex network is an emerging hot topic in modern scientific research.Complex network and social reality has a lot of close contact, there are many complexnetwork models in the real world, for example: social networks, species prey relationshipsnetworks, communication network and so on. It is an issue of great concern to scientistsand researchers that applied to other areas of the complex network model to solve somepractical problems in the society. The findings of a large number of scholars, most complexnetworks have some of the same characteristics, such as shorter average network path, thelarger the network clustering coefficient, the power law distribution of node degrees. Thestructure and characteristics of the network model of processing real network has a guidingrole. This article focuses on how to solve TSP problems on the background of a complexnetwork of knowledge. First, the article describes the basic idea of the TSP. Second,analysis of the traditional intelligent algorithm for solving TSP, elaborate the geneticalgorithm, ant colony algorithm and particle swarm optimization algorithm description,conduct simulation experiments, and analysis of the experimental results. Third, elaboratethe concept of a complex network and its models, give the main statistical characteristics ofthe complex network, mainly including the degree distribution, clustering coefficient andaverage path length. The paper also gives a detailed modeling algorithm of small-worldnetworks and scale-free networks, and WS, NW and BA three network model simulationexperiments. Finally, design a specific algorithm to solve TSP problem based on thecomplex network, write codes to simulate experiment, a more in-depth analysis of theexperimental results, and analysis of the experimental results. The experiments proved theeffectiveness and feasibility of the algorithm.