基于案例推理的指挥实体任务规划高维解空间适应性问题研究∗

摘要

Adaptation is the most difficult stage in the Case⁃Based Reasoning cycle, especially, when the solution space is multi⁃dimensional in the Command Entity�s Mission Planning. This paper discusses the adaptation of a high dimensional solu⁃tion space in the Command Entity�s Mission Planning and proposes a possible approach to it. A Visualization induced Self Or⁃ganizing Map is used to map the problem space and solution space first, then a Back Propagation network is applied to ana⁃lyze the relations between these two maps. A simple military scenario is used as a case study for evaluation purposes.%利用案例推理对指挥实体任务规划过程中决策问题求解方法的修正过程是该方法推理过程中最困难的阶段，尤其当决策问题解空间是多维的情况下。讨论了指挥实体任务规划过程中高维决策空间的修正问题，并提出了可行的解决方法。首先，利用自组织匹配法清晰展现问题空间与决策空间的映射过程，然后，利用反向传播神经网络分析匹配结果间的相关性问题，最后，选取一个简化的军事剧情对该方法的合理性进行验证。