中文摘要：

本文根据影响并行蚁群算法性能的关键因素，提出了一种自适应的并行蚁群算法．首先提出了基于适应度和基于距离选择的两种不同的信息交流策略，使得各处理机自适应地选择与之进行信息交换的处理机，然后采用自适应的更新策略进行信息素的更新．为了增强该算法的搜索能力，还根据解的多样性给出了自适应地调节处理机之间的信息交流周期的方法．在MPP处理机深腾1800上对TSP问题的实验结果表明了该算法在保证有效的加速比的同时，具有很好的收敛性．

英文摘要：

An adaptive parallel ant colony algorithm is presented by considering the critical factors influencing the parallelization of the ant colony algorithm. Two different strategies for information exchange between processors are proposed： selecting the partner processor based on the fitness and on the distance. These strategies enable each processor to choose the partner proces- sor to communicate and update the pheromone adaptively. In order to increase the ability of optimization and to avoid early convergence, we also propose a method of adjusting the iteration interval adaptively according to the diversity of the solutions. These techniques are tested on the traveling salesman problem using the massive parallel processors （MPP） Shengteng1800. Experimental results show that our algorithm can converge very quickly and has high speedup.