Abstract:Abstract Summarizing the current situation of knowledge engineering languages and their support environments. We designed and implemented the NEW expert system development environment according to the idea that knowledge engineering languages and their environments are the evolution of traditional programming. This paper describes its design and implementation,focusing upon the architecture,the knowledge representation,the inference mechanism,the environment and the applications.

Abstract:The reasoning with fuzzy quantifiers has much importance in computer science. In this paper we deal with fuzzy quantifiers and reasoning from the view point that fuzzy quantifiers have statistical property. The results may have applications in expert systems and other intelligence systems.

Abstract:CPANEL Plus is an advanced software for designing and managing screen panels. In this paper, we will introduce its purpose, basic components, Chineselization technique examples and applic tion prospects.

Abstract:In the case of region edge recognition when several edges intersect or there are some edges which belong to several different regions, the algorithm REMR (Recognition of the Edges of the Minimal Region) can find out the whole edges of the minimal region by line tracing. Then algorithm EBRM ( Edge-Based Region Mark ) marks the region with the recognized edges. These two algorithms have been applied to the image preprocessing system for computer-controlled embroidery.

Abstract:Knowledge refinement is an important aspect of knowledge acquisition and is used to solve the problem of correctness and efficiency of knowledge base. This paper discusses the basic theory and method of knowledge refinement through the introduction of implementing method of refiner KBRS in the knowledge base constructing tool GKD-KBST.

Abstract:In order to solve the problem of KB s incompleteness in Explanation-Based Learning. We introduce Multi-KB cooperation in EBL and propose a new learning method-Explanation and Association-Based Learning.

Abstract:Based on the parallel computation model Petri Net, we ll introduce a kind of Ordered Petri Nets, whose computing power is much more stronger compared with Petri Net, by giving a partial order on the transition set of Petri Net, and define a parallel model of computing machine: Ordered Petri Net Computing Machine. In the last section, it is proved that this kind of machine can compute all the general recursive functions, and so it can be regarded as the basic model of computing machine.

Abstract:According to the idea of categorical equivalence, we discuss a categorical equivalence on Horn logic. The method of categorical equivalence could be treaded as a method to describe semantics of meta-level operation. We generalize the method.

Abstract:In this paper,we describe the deviation of confidence factor about an assertion,the rule intensity deviation, the assertion reliability (the reliable degree about an assertion), the rule reliability (the reliable degree of a rule) ,and the level two uncertainty. Based on these concepts we present an inference method that can deal with the deviation of confidence factor about assertions and the rule intensity deviations,a new reasoning model which integrates the two levels of uncertainty (the confidence factor and the reliable degree).Adiscussion about the evaluation of conclusions given by the cooperative ESs is also in progress in this paper.

Abstract:This paper purports to present an assertional proof for Peterson-Fischer 2-mutual exclusion algorithm. According to Nancy A. Lynch of MIT[1], this was an open problem. Hence the significance of this paper is to fill the gap.

Abstract:To study the Non-INF relational model, G. M. Kuper proposed the LPS language. Based the LPS language,a procedural semantics with mgu s are introduced ,and we implamented the LPS language using meta-interpreter method.

Abstract:A new and practical algorithm for FORTRAN parallelism detection is described. In this paper,a array-dependence-on-loop concept that is important for detecting parallelism is presented first. Some new theories of detecting parallelism as well as a practical decision algorithm are then given.