改进的Tabu Machine网络求解最大割问题

摘要

In order to improve the solution quality of maximum cut problem processed with Tabu Machine Network, by improving the tabu search strategy of previous algorithm and combining local search strategy with estimation of distribution strategy, a new network, which is called HNNTS-EDA for short, is created. This network has preferably stronger local search ability and the capacity of getting out of the local optimum value. By comparing the HNNTS-EDA network with a number of classic algorithms on testing the same data, it is proved that it possesses stronger search ability.%为了提高Tabu Machine网络处理最大割问题时解的质量,改进原有算法的禁忌搜索策略,并且通过结合局部搜索策略和分布估计策略,形成一种新的网络HNNTS-EDA.此网络有较强的局部搜索能力和脱离局部最优值的能力.将HNNTS-EDA网络与多种经典算法在相同测试数据上进行对比测试,实验结果表明HNNTS-EDA网络具有更强的寻优能力.