一种改进的并行蚁群优化求解算法

An Improved Parallel Ant Colony Optimization Zlgorithm

为了提高并行蚁群优化算法的求解性能,对ACO算法进行了改进.针对有明显聚类特征的大规模TSP问题,充分利用问题本身所具有的特征,提出了一种带聚类处理的蚁群算法,该算法比较ACS算法可以在更短的时间内找到相同质量的解,而且在相同的运行时间内,该改进算法总能找到最好的解.在VC++环境下进行仿真实验,求解了TSP库中的实例pr136、pr107,分别得到了其最短距离,结果表明了编程思路的正确性及高效性.

In order to impoove the capability of parallel ant colony algorithm（ACA）,ACO is improved. To tackle large-scale traveling salesman problem（TSP） with characteristic of clear clustering,a new ACA algorithm is proposed by making full use of the issue itself which has the characteristics. Comparing ACS this improved algorithm cannot only find the same quality of the solution within a shorter period of time but also obtain the best solution within the same time. The simulation in the VC＋＋ environment that applied the TSP（traveling salesman problem） about 136 cities and 107 cities reache two shortest distance. The results show that the way in programing is correct, and the algorithm is efficient.