[1]苏守宝,赵威,李智.求解加权MTSP问题的CUDA并行群智能方法[J].郑州大学学报(工学版),2021,42(06):35-42.
点击复制

求解加权MTSP问题的CUDA并行群智能方法()
分享到:

《郑州大学学报(工学版)》[ISSN:1671-6833/CN:41-1339/T]

卷:
42
期数:
2021年06期
页码:
35-42
栏目:
出版日期:
2021-11-10

文章信息/Info

作者:
苏守宝赵威李智
文献标志码:
A
摘要:
针对使用混合迭代启发式算法求解多旅行商问题(MTSP)执行速度慢的问题,根据群智能算法具有良好的并行特性,对粒子群算法和蚁群算法在GPU上并行化实现和编程优化等方面进行相关研究后,提出一种基于CUDA的混合粒子群聚类-蚁群算法(GPSO-AC)。该算法利用GPU具有多个流处理器和单指令多数据(SIMT)的架构特点,将算法中的大量独立个体的搜索过程同时执行。实验表明,该算法执行速度远快于CPU版算法且随着问题规模扩大速度差距越明显,在收敛精度上也优于同类算法,同时探讨了加入代价均衡约束后对含权MTSP问题最优解收敛精度的影响。
Abstract:
To solve the low running speed of the multi-traveling salesman problem(MSTP) using the heuristic method,a CUDA-based hybrid particle swarm clustering-ant colony algorithm (GPSO-AC) was proposed by integrating their parallel characteristices with programming techniques optimally.GPSO-AC used GPU’s instruction architecture with multiple stream processors(SM) and single instruction multithreading (SIMT)  to parallel the search process of numerous independent individuals, so as to accelerate the execution speed of the hybrid iterative method. GPSO-AC was tested on 6 datasets compared with other methods, such as PSO-AC, TPHA and K-means-AC. Then the influence of cost equilibrium constraint on the convergence performance of the optional solution of weighted MTSP problem was dicussed. Furthemore, the cost standard deviations obtained from GPSO-AC on chn31 with different traveling salesmen, were 1165.26,54,97 and 6.74 in the three cases respectively. The experimental results showed that the proposed algorithm was much faster than other cases CPU based algorithms and the advantage becomed more obvious with the expansion of the model size, and the convergence precison of the algorithm was better than the similar algorithms for solving MTSP problems
更新日期/Last Update: 2021-12-17