论文部分内容阅读
针对量子粒子群优化算法面对复杂优化问题时,临近最优解的搜索阶段存在收敛速度慢、在边界附近全局搜索性差的问题,提出了基于CUDA的边界变异量子粒子群优化算法.GPU(图形处理器)以多颗密集的计算核心模拟粒子的搜索过程,利用并发的优势提升粒子搜索速度;边界变异则通过以随机概率将边界粒子扩散到更大的搜索域,增加种群的多样性,提升粒子群的全局搜索性.对若干优化算法的仿真实验表明,所提出方法具有较好的全局收敛性,且同等目标精度下,取得了较高的有效加速比.
In order to solve the problem of complex optimization in quantum particle swarm optimization (PSO), the problem of slow convergence and global search near the boundary in the search phase near the optimal solution is proposed, and a CUDA-based boundary-variant quantum particle swarm optimization algorithm is proposed. Processor) to simulate the particle search process with multiple dense computing cores and use the advantages of concurrency to enhance the particle search speed. Boundary variation increases the diversity of the population by increasing the particle size to a larger search domain with random probability The global search of particle swarm optimization experiments on several optimization algorithms show that the proposed method has better global convergence and a higher effective speedup under the same target precision.