Theory System Construction of Distance-Type Fuzzy Reasoning Method

本课题旨在构建距离型模糊推理方法的理论体系，以实现类人推理。包括去年的成果在内，以下项目已经完成。人类不仅使用单阶段多重规则，而且使用多重多阶段规则来完成高级推理。因此，去年的基本推理算法被扩展到多重多阶段规则。其基本思想是，将规则中模糊变量之间的距离视为电阻，根据电网络理论，可以推导出一种具有离散模糊变量的多重多级规则的推理算法.在脑内的知识表征中，视觉图形信息是形象化的，由于信息量大，不仅符号信息而且图形信息也很重要。因此，知识的图形表示常用于推理。为了实现更人性化的推理，引入特征空间和概念向量的表达，将图形信息作为模糊集的扩展，提出了距离型图推理方法.这是可取的讨论知识表示，知识获取，学习和推理集体。因此，在以前的研究中，我们提出了一个算法，所有的智能活动与推理方法，满足距离概念和分离规则。仿真结果表明了该方法的有效性，并在此基础上开发了一个人机推理引擎软件。计划通过日本模糊理论和智能信息学会的主页分发该软件。

In this project, theory system of Distance-Type Fuzzy Reasoning Method is being constructed to realize humanlike reasoning. The following items have been completed including last year's results.1. Humans use not only single-stage multi-fold rules but also multi-fold multi-stage rules to fulfill high level reasoning. Therefore, last year's basic reasoning algorithm was expanded to multi-fold multi-stage rules. The basic idea is that if the distance between fuzzy variables of rules is regarded as resistance, according to the theory of electric network, it is possible to develop a reasoning algorithm for multi-fold multi-stage rules which have discrete fuzzy variable.2. In the knowledge representation within the brain, visual figure information is figurative, and since there is much amount of information, not only symbolic information but also figure information is very important. Therefore, the knowledge representation with figures is often used for reasoning. In order to realize more humanlike reasoning, Distance-type Diagrammatic Reasoning Method has been developed by introducing the expression of feature space and conceptual vector, regarding figure information as the expansion of fuzzy set.3. It is desirable to discuss knowledge representation, knowledge acquisition, learning and reasoning collectively. Accordingly, in previous studies, we have proposed an algorithm to unite all the intelligent activities with a reasoning method satisfying distance concept and rule of detachment. Its effectiveness has been shown at simulation level.A software named Human Reasoning Engine has been developed based on the above-mentioned results. There is plan to distribute the software through the homep age of Japan Society for Fuzzy Theory and Intelligent Informatics.