基于广义三维分配的数据互联算法研究

Study on Generalized 3-D Distribution Based Data Association Algorithm

针对多维分配问题中基于次梯度算法的拉格朗日松弛算法每次迭代都要对每个子问题进行最小化运算以更新乘子的缺陷,引入了优化数学里的代理次梯度算法,修改次梯度表达式和乘子更新公式,提出了基于代理次梯度的拉格朗日松弛数据关联算法。在问题规模较大的情况下,节约了计算时间,降低了跟踪丢失率。仿真结果证明了算法的有效性。

Aimed at the defect that general sub gradient based on Lagrangian relaxation algorithm of multidimensional distribution problem needs to minimize all sub problems at every iterative time, a surrogate sub gradient optimization mathematical algorithm was introduced. The sub gradient was modified as well as multipliers updating expression. And a surrogate sub gradient based on Lagrangian relaxation data association algorithm was proposed to save time and reduce track loss in a large scale problem, The result of simulation proves its efficiency.