并行蚁群算法在虚拟场景下寻路中的应用

The Application of Parallel Ant Colony Algorithm in Virtual Scene Path-Finding

本文根据蚁群算法的并行特性,对并行集群下的同步异步蚁群算法进行研究,阐述了在TSP问题求解中的仿真应用,通过实验对并行蚁群算法的参数选择进行了分析,确定了参数的选择原则以及对算法性能的影响,并且应用于虚拟场景下车辆的寻路,有利于并行蚁群算法在优化问题中的推广和应用。仿真研究表明并行蚁群算法不仅具有较快的寻优速度,而且具有较强的寻优性能。该算法中的参数对于TSP的求解有较大的影响,通过最优参数的正确选取,能使得算法取得更优的值。

Considering the parallelization of the ant colony algorithm, we do our research on Synchronous / asynchronous ant algorithm under parallel cluster platform, give its application in TSP,and analyse the selection of parameters impacting on the performance of ACS. The results from this paper are beneficial to the application and development of the ant colony algorithm in optimization problems, and apply to path-finding in virtual scene.. Simulation experiment shows that the parallel method not only has a comparatively high speed on optimization, but also have a higher performance on optimization. Also, the parameters of ACS algorithm under parallel cluster platform have the very big influnence to the TSP, by selecting the Optimal Parameters rightly, it can makes the algorithm to reach the optimal value.