启发式优化算法的GPU并行加速框架

GPU Parallel Acceleration Framework for Heuristic Optimization Algorithm

为解决启发式优化算法计算量大、耗时长的缺点,使用图形处理单元(GPU)以及统一计算架构(compute unified device architecture,CUDA)对启发式优化算法进行并行化。提出了一种针对启发式优化算法的GPU并行框架,设计了具有并行逻辑结构的信息交互框架、算法并行优化策略,解决了信息交互的逻辑结构在串、并行中的相异性问题,该框架可并行化各类启发式优化算法,具有一般性与高效性。为验证该框架的有效性,利用并行框架对5种常见启发式优化算法进行并行化,给出了多个测试函数下GPU并行计算与CPU串行计算的对比结果,其中差分进化算法、哈里斯鹰优化算法、灰狼优化算法、鲸鱼优化算法在种群维度为5000时,分别加速高达179.1、178.6、74.3、358.2倍,同时保证了结果的准确性,表明所设计并行框架的高效性与实用性。

Heuristic optimization algorithm are a type of algorithm that uses large-scale populations for iterative calculations and are widely used to solve all kinds of complex optimization problems.However,such algorithm have the disadvantages of large calculation and long time consumption.To solve this problem,heuristic optimization algorithms are parallelized using GPU and compute unified device architecture(CUDA)to substantially improve computational efficiency.A GPU parallel framework for heuristic optimization algorithm is proposed,which designs an information interaction framework and algorithm parallel optimization strategy with a parallel logical structure,and solves the problem of the dissimilarity of the logical structure of information interaction in series and parallel,this framework can parallelize various heuristic optimization algorithms with generality and efficiency.In order to verify the effectiveness of this framework,five common heuristic optimization algorithms are parallelized by using the parallel framework,and the comparison results of GPU parallel computation and CPU serial computation under different multiple test functions are given.in which DE,HHO,GWO,and WOA reach the acceleration ratio of 179.1,178.6,74.3 and 358.2 times respectively when the population dimension is 5000,while ensuring the accuracy of the results,which verifies the high effectiveness and practicability of the designed parallel framework.