摘要： 针对虚拟作战个体行为具有智能性和自主性的特点，从行为功能考虑，将影响人员外在行为的内部因素分解为元Agent，提出一种基于元Agent聚集的作战人员行为模型。依据各个主体特点采用智能模型实现其结构，描述感知因素的量化值和可信度，引入收益指标评估意图的远景规划，通过动作执行概率的赋值达到重复作战行为的优先选择。与传统方法相比，该模型充分考虑到个体行为内在因素间的交互耦合和复杂反馈，提高了作战个体战场行为的智能性、自主性和差异性，为作战人员行为建模等问题提供一种新的研究思路。

关键词: 元Agent, 体系结构, 行为建模, 作战个体

Abstract: Aiming at actions of virtual battlefield individual with the intelligence and autonomy characteristic, internal factors that influence personnel external behavior are broken down into meta-Agents with considering behavior functions, and a battlefield individual behavioral model is advanced by aggregating meta-Agents. Meta-Agent structure is accomplished by intelligent model based on its main features, with describing values and reliability of perception factors, evaluating the vision of intention by revenue target, carrying out repeat battlefield action preferential through the probability of actions performed. Compared with traditional methods, the meta-Agents aggregation behavioral model greatly improves the intelligence, autonomy and diversity of behavior by handling the coupling and complex interactions of meta-Agents, which provides a new approach to battlefield individual action modeling.

Key words: meta-Agent, architecture, behavior modeling, battlefield individual

中图分类号: 

• TP18