基于小样本数据的作战计划生成方法

Combat Project Creating Based on Small Quantity Data

针对部队实兵对抗实验数据生成和采集的投入高、难度大的实际,提出基于小样本数据生成作战计划的方法:首先,采用偏最小二乘回归分析法建立作战因素与作战结果间的多元非线性模型;其次,采用改进的遗传算法,在MATLAB平台编程实现了对模型的整数规划求解,并通过实例验证算法的有效性;最后,利用作战因素的重要性排序分析作战计划的调整流程。实例分析表明,该方法具有较好的应用价值。

The data creating and collection is costly and difficult in army＇s veritable oppositional experiment.So,the paper bring forward the method of creating combat plan based on a small quantity of data swatch：Firstly,use the method of partial least-squares to establish the non-linearity model for the combat factors and the result.Secondly,use the reconstructive genetic algorithm to calculate the integer programming for the model on the MATLAB system,and validate the algorithm by an example.Lastly,to research the redressal procedure for the plan by analysing the importance sequence of combat factors.