针对复杂战场环境中的意图推理问题,首先构建能够表示目标信息及相互间关系的基于多智能体的意图推理网络,将目标意图推理问题转化为复杂网络中的最优推理路径搜索问题;然后,在基本人工蜂群算法的基础上,引入遗传算法中的遗传变异交叉等思想,并结合改进的食物源自适应策略,提出一种自适应遗传蜂群推理算法;最后,通过实验对比,验证了自适应遗传蜂群推理算法在多智能体网络推理中具有比其他算法更优的准确性和收敛性。

In view of the intention reasoning problem in complex battlefield environment, an intention reasoning method based on multi-agent network is established, which can indicate the object information and interrelationship, and the object intention reasoning problem is transformed into the reasoning path search optimization in complex networks. Then, besides the improved food source adaptive strategy, the cross mutation is introduced on the basis of the basic artificial bee colony algorithm, an adaptive genetic bee colony reasoning algorithm is proposed. Finally, through experimental comparison, it is verified that the intention reasoning method by adaptive genetic bee colony alogorithm gains better accuracy and convergence than other multi-agent network inference algorithms.

E91;E926.4