首页> 外文会议>International Conference on Information Fusion >Large-scale Multi-dimensional Assignment: Problem Formulations and GPU Accelerated Solutions
【24h】

Large-scale Multi-dimensional Assignment: Problem Formulations and GPU Accelerated Solutions

机译:大规模多维分配:问题配方和GPU加速解决方案

获取原文

摘要

In this paper, we present alternate integer programming formulations for the multi-dimensional assignment problem, which is typically employed for multi-sensor fusion, multi-target tracking (MTT) or data association in general. The first formulation is the Axial Multidimensional Assignment Problem with Decomposable Costs (MDADC). The decomposable costs comes from the fact that there are only pairwise costs between stages or scans of a target tracking problem or corpuses of a data association context. The difficulty with this formulation is the large number of transitivity or triangularity constraints that ensure if entity $A$ is associated to entity $B$ and entity $B$ is associated with entity $C$, then it must also be that entity $A$ is associated to entity $C$. The second formulation uses both pairs and triplets of observations, which offer more accurate representation for kinematic tracking of targets. This formulation avoids the large number of transitivity constraints but significantly increases the number of variables due to triples. Solution to large-scale problems has alluded researchers and practitioners alike. We present solution methods based on Lagrangian Relaxation and massively parallel algorithms that are implemented on Graphics Processing Units (GPUs). We test the problem formulations and solution algorithms on MTT problems. The triples formulation tends to be more accurate for tracking measures and the MDADC solver can solve much larger problems in reasonable computational time.
机译:在本文中,我们为多维分配问题提供了备用整数编程配方,其通常用于多传感器融合,多目标跟踪(MTT)或数据关联。第一种配方是具有可分解成本(MDADC)的轴向多维分配问题。可分解的成本来自于目标跟踪问题的阶段或扫描之间只有成对成本或数据关联上下文的语料。这种配方的难度是大量的传递性或三角形限制,以确保实体 $ a $ 与实体相关联 $ b $ 和实体 $ b $ 与实体有关 $ c $ ,那么它也必须是那个实体 $ a $ 与实体相关联 $ c $ 。第二种配方使用对和三胞胎观察,为目标的运动跟踪提供更准确的表示。该配方避免了大量的传递约束,但显着增加了Triples引起的变量的数量。大规模问题的解决方案已经暗示了研究人员和从业者。我们提出了基于拉格朗日放松的解决方案方法,并在图形处理单元(GPU)上实现的大规模并行算法。我们在MTT问题上测试问题配方和解决方案算法。 Triples制剂往往更准确地用于跟踪措施,并且MDADC求解器可以在合理的计算时间内解决大量问题。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号