单炮多发同时弹着的局部插值算法与并行加速

为提高单炮多发同时弹着的弹道计算精度,基于标准条件下二维质点弹道方程和真空二维弹道斜射程公式,提出了一种局部插值算法。通过缩小插值区间加密插值序列提高弹道计算精度,研究了该局部插值算法的MPI并行实现方法,该方法可有效减少弹道计算消耗的时间,保证计算耗时满足战场要求。给出了并行计算的任务分配方案和并行效率的计算。并行效率与计算核心数量的关系表明弹道计算机的CPU核心数存在一个上限。以5~15km射程的同时弹着弹道计算为例,验证了该文算法的有效性和高效性。

Local Interpolation Algorithm and Parallel Computation of Multiple-round Simultaneous-impact of Single Artillery

In order to improve the computation accuracy of the ballistics for multiple-round simultaneous-impact( MRSI) of single artillery, the local interpolation algorithm ( LIA) was proposed based on 2D particle trajectory equation under standard conditions and vacuum 2D trajectory oblique-range formula. It can obviously improve the accuracy to reduce the interpolation interval and increase interpolation sequence. The MPI based parallel computation technique was used in the local interpolation algorithm. The method can effectively reduce the computation time and ensure the computation time satisfy battlefield requirements. The task allocation scheme in the parallel computation was given,and the parellel efficiency was computed. Relationship between this parallel computation efficiency and the number of CPU cores used shows that there is an upper limit on number of CPU cores of ballistic computer. The MRSI computa-tion task was carried out with range between 5 km to 15 km to verify the validity and efficiency of the algorithm.