基于规则提升的Agent系统异常处理方法

基于进化计算理论中的重组思想, 提出一种基于规则提升的Agent系统异常处理方法, 处理Agent系统中违背执行规则的异常. 该方法根据规则间的相互触发和执行情况分析得到规则处于异常状态的原因, 重组处于异常状态的规则, 提升规则向适应环境的方向进化. 实验结果表明, 该方法能处理规则的异常状态, 增强了规则的触发条件, 保证了规则达到预期的后置条件, 提高了Agent系统在开放环境下的异常处理能力.

Exception Handling Method of Agent System Based on Norm Improvement

Based on the idea of recombination of evolution computing theory, we proposed an exception handling method of an Agent system based onnorm improvement to deal with the exceptions that violated execution norms in Agent system. The proposed method could analyze the reason for a norm exception based on the mutual triggering and execution information among norms, recombine the norms that were in the exception state, and improve the norm to make it more adaptable to the environment. Theexperimental results show that the method can deal with the exception state of the norm, enhance the triggering condition of norms, ensure that the normmeet the expected postcondition, and improve the ability of exception handling of Agent system in an open environment.