Hopfied神经网络在TSP问题中的应用-应用数学专业毕业论文.docx

中北大学学位论文之间，说明该方法对于处理旅游路线的选择问题是行之有效的。 中北大学学位论文 之间，说明该方法对于处理旅游路线的选择问题是行之有效的。 关键词：旅行商，组合优化，遗传算法，模拟退火算法，蚁群算法 中北大学学位论文The 中北大学学位论文 The application of Hopfield neural network in the TSP problem Lan Zhaoqing Tutor：Bai YanpingAbstract TSP is a typical problem in the field of combinatorial optimization，the core of this problem is to find the shortcut including all the cities．Although it is very simple to state，it is not SO easy to solve，what is more，it has been proved to be the NP—complete problem．But it does exist extensively，and it is the central generalize and simplified form of many complex issues．Therefore propose an effective solution to the problem of the TSP algorithms have a higher theoretical and practical value． This paper chose the existing algorithms of TSP to start．By studying a large number of references；to understand the main ideas on variety 16f algori衄霹葡d organize a variety of algorithms，then make the classification．It is found that genetic algorithm，simulated armealing algorithm and ants algorithm showed a certain advantages in solving the TSP，and they are widely used in the solution of practical problems．Then，to this paper，the three algorithms is deeply researched and programmed．The three algorithms were separately used to solve the 4≤ cities TSP problems．From the results of the simulated，the author found that annealing’s process of optimization is longer；ant algorithm is also relatively long，but also easy in a local optimal solution to search stagnation；At the practical application of premature，GA appears the shortcomings of easy to premature convergence and convergence of poor．How fast and accurate to solve the problem has now become a difficult point in the problem of TSP algorithm．The author presents a Hopfield neural network algorithm，the neural net work is the parallel computing，and its calculation is not the dimension of the increase in the index of ”explosion”and therefore,the optimization of high—speed computing particularly effective．In the actu