基于GPU加速遗传算法的直接定位研究

针对大规模数据下遗传直接定位算法执行时间慢、实时性较差的问题,提出了基于GPU加速的并行遗传直接定位算法。根据直接定位代价函数特点,设计了GPU高速并行遗传进化架构,通过对适应度函数并行化计算以及对选择、交叉、变异等遗传操作并行化设计,缩短了算法执行时间,提高了算法执行效率。仿真实验表明,通过合理的GPU并行线程结构设计,显著提升了遗传直接定位算法的执行速度,可更快得到直接定位估计结果。

Research on direct position determination based on GPU accelerated genetic algorithm

The genetic direct position determination (DPD) algorithm executes slowly and has poor real-time performance under the large-scale data condition. This paper proposed a GPU-based genetic DPD algorithm to overcome the above shortcoming. According to the cost function of DPD, it designed a high speed parallel architecture of GPU. It reduced the execution time and improved efficiency, according to the parallel design of the fitness function and the genetic operation such as selection, crossover and mutation. The experiments show that, through reasonable design of the parallel thread architecture of GPU, the proposed method can reduce the execution time of the genetic DPD algorithm efficiently and locate the emitter faster.