2.118

影响因子

    高级检索

    基于KG+LLM的联合作战计划智能生成方法

    Intelligent Generation of Joint Operation Plans Based on KG+LLM

    • 摘要: 当前国内外对于联合作战计划生成通常遵循基于预案的 “快速修案” 和基于操作流程与作战规则的 “慢速制案” 模式,但由于在制案过程中作战信息未成体系化管理、信息交互方式繁琐、人为理解存在偏差等,导致制案周期较长,难以满足实际作战需要,因此,提出了基于KG+LLM的联合作战计划智能生成框架,通过知识图谱和大模型技术缩短了制案周期,结合 “安东诺夫机场闪击战” 设计了基于KG+LLM的人机协作 “智能拟案” 模式,在CMO海空兵棋系统验证了框架的可行性和有效性。

       

      Abstract: Joint operations plan generation, both domestic and international practices predominantly adhere to two models: the“rapid revision of plan”approach grounded in pre-existing plans, and the“slow plan-making”method rooted in operational procedures and combat rules. However, the process of plan making is often encumbered by disorganized combat information management, a complex information exchange process, and cognitive biases among personnel, collectively leading to protracted planning cycles that fall short of meeting the exigencies of actual combat scenarios. Consequently, framework for intelligent generation of joint operations plans based on KG+LLM is proposed. The period of plan making is shortened by knowledge graphs and large models technologies. Furthermore, the “Antonov Airport Blitzkrieg”scenario is integrated to design a human-machine collaborative“intelligent drafting”model based on KG+LLM. The feasibility and viability of this framework have been demonstrated through its application in the CMO sea-air wargaming system.

       

    /

    返回文章
    返回