摘要
针对复杂人机交互实时性的要求,提出了一种基于优化算子的SIMD并行碰撞检测算法。引入了优化算子,将搜索空间限定在非均匀的局部极小区域,减少了蚁群的搜索时间。在多蚁群求解过程中,将子任务使用负载均衡策略分配到多核处理器的各个处理核心上并行执行,实验结果表明:与经典的I-COLLIDE、MPI及Pipelining等算法相比,本文提出的算法较好地解决了人机交互中的碰撞检测问题。
To meet the real-time requirement in complex human-computer interaction,a Single Instruction,Multiple Data(SIMD)parallel collision detection algorithm based on optimization operator is proposed.By introduction of optimization operator,the search space is confined in a nonuniform local minimum area,reducing the colony search time.In the process of solving the multiple ant colony,the load balancing strategy is used to assign the sub-tasks to each processing core on multi-core processors for parallel execution.Results show that,compared with the classic ICOLLIDE,MPI and Pipelining algorithms,the proposed algorithm has better performance in solving the human-computer interaction in collision detection.
出处
《吉林大学学报(工学版)》
EI
CAS
CSCD
北大核心
2017年第5期1598-1603,共6页
Journal of Jilin University:Engineering and Technology Edition
基金
吉林省自然科学基金面上重点项目(20140101196JC)
浙江省自然科学基金面上项目(LY15F020017
LY15F020016)
吉林省教育厅"十三五"科学技术研究重点项目(吉教科合字[2016]第174号
[2017]JJKH20170990KJ)
关键词
人工智能
碰撞检测
并行算法
优化算子
平衡包围盒
artificial intelligence
collision detection
parallel algorithm
optimization operator
balance bounding box