论文部分内容阅读
在国内的综合建模形式弹道极限方程中存在11个待定参数,从理论上讲,采穷举法可以获得其数值大小,但需要的计算时间过长和储存空间巨大,不宜实现,为解决此问题,改用差异演化算法。基于国内填充式实验数据,采用差异演化算法对综合建模形式弹道极限方程的11个待定参数进行了多目标优化计算,结果显示,方程的总体预测率为82.35%、安全预测率为100%、平均相对误差平方和为0.0013。该方程对其他来源的49个实验数据的预测结果显示总体预测率提升了1.32%,安全预测率降低了4.08%,平均相对误差平方和增加了0.0073,表明差异演化算法适用于解决多参数多目标的弹道极限方程建模问题。
In our country, there are 11 undetermined parameters in the ballistic limit equation of integrated modeling. Theoretically speaking, the method of taking pluton can obtain the numerical value, but the calculation time is too long and the storage space is huge, so it is not suitable to solve this problem Problem, use differential evolution algorithm. Based on the domestic experimental data, the multi-objective optimization of eleven unknown parameters of the integrated modeling ballistic limit equation is carried out using the differential evolution algorithm. The results show that the overall prediction rate of the equation is 82.35% and the safety prediction rate is 100% The sum of squares of the average relative error is 0.0013. The prediction of the 49 experimental data from other sources shows that the overall prediction rate is improved by 1.32%, the safety prediction rate is reduced by 4.08% and the mean square error of the relative error is increased by 0.0073, which shows that the differential evolution algorithm is suitable for solving multi-parameter and multi-objective Ballistic Limit Equation modeling problem.