基于人工智能的多Agent协同辩证逻辑推理方法  

Multi-Agent Cooperative Dialectical Logic Reasoning Method Based on Artificial Intelligence

作　　者：黄海 桂起权[3] Hai Huang;Qiquan Gui

机构地区：[1]许昌学院马克思主义学院 [2]同济大学移动源后处理研究院 [3]武汉大学哲学院

出　　处：《逻辑学研究》2023年第5期81-96,共16页Studies in Logic

摘　　要：为消除Agent内部信念冲突,实现多Agent的协同运行,利用人工智能技术,设计多Agent协同的论辩式辩证逻辑推理方法。根据Agent的组成结构以及多Agent之间的通信方式,建立多Agent协同模型。在该模型下,利用人工智能技术,求解多Agent协同任务,消解Agent间冲突。通过计算识别出微小变化会导致系统行为重大变化的关键变量,找到推理依据,在推理论据和规则的支持下,实现多Agent协同的论辩式辩证逻辑推理。实验结果表明,设计论辩式辩证逻辑推理方法的加速率提高了1.4,多Agent之间的协同程度提高了4.12,同时抵制恶意节点的成功率提升了12.6%,具有更好的应用效果。In order to eliminate the conflict of beliefs within agent and realize the cooperative operation of multiple agents,a multi-agent collaborative dialectical logic reasoning method is designed by using artificial intelligence technology.According to the composition structure of agent and the communication mode between multiple agents,a multiagent collaboration model is established.Under this model,artificial intelligence technology is used to solve multi-agent collaborative tasks and resolve conflicts between agents.By calculating the critical factor and critical element,the dialectical logic inference evidence of multi-agent collaboration is constructed,and with the support of inference evidence and rules,the multi-agent collaborative dialectical logic reasoning is realized.The experimental results show that the acceleration rate of designing the dialectical logic reasoning method is increased by 1.4,the degree of cooperation between multiple agents is increased by 4.12,and the success rate of resisting malicious nodes is increased by 12.6%,which has a better application effect.

关 键 词：人工智能 多AGENT协同 论辩式辩证逻辑 逻辑推理 

分 类 号：B81[哲学宗教—逻辑学]