农业HPC专家系统不确定性推理机的设计

A Uncertain Reasoning Machine

讨论了农业HPC专家系统的不确定性推理机的知识表达形式、可信度定义、证据不确定性的表示、组合证据不确定性的计算以及不确定性的更新等问题 ,给出了该不确定性推理机的控制策略 ,该策略的主要特点是采用一种目标引导的逆向推理 ,并结合以可信度为基础的冲突消解策略。同时还具体介绍了推理机的算法流程 ,最后以葡萄营养失调的诊断决策为例 。

This paper firstly discusses some problem about the uncertain reasoning machine in the agricultural expert system based on HPC, such as the knowledge representation, the certainty factor, the representation of evidence uncertainty, the calculation of multi evidence uncertainty, the update of uncertainty and so on. Secondly, the control strategy of this uncertain reasoning machine is discussed. This control strategy adopts a converse reasoning strategy based on target, and it adopts a conflict removing strategy based on certainty factor. The arithmetic to realize the uncertain reasoning machine is also given in detail. At last, this paper stats the example of malnutrition diagnosing of grape as the application of the reasoning machine.