Computer-aided software engineering of expert systems

Evaluating the benefits of a computer-aided software engineering tool to develop and document product configuration systems

“…Accordingly, the PCS challenges, available tools, the relations between the challenges [8,[21][22][23][24]27,29], and the potential solutions have been listed [1,19,28,49,51,57,62]. Moreover, we build the suggested CASE tool snippets based on the previously proposed solutions in the literature as well as the reported benefits from the CASE tools, as listed in Table 1 [33,[59][60][61]. Hence, the goal of the proposed tool is to combine, integrate, and mainly automate model-driven development of PCS software based on the PVM and CRC cards while minimising the reported challenges [33,[59][60][61].…”

“…Standardisation of system development methodologies (automatic performance) [60,61] One of the most important components of CASE tools is an extensive data dictionary, which keeps track of all objects created by the system designer. Data dictionary [60] The CASE data dictionary stores data flow diagrams, structure charts, descriptions of all external and internal entities, data stores, data items, report formats, and screen formats.…”

ALCA — automated local area networks configuration aid

“…While embarking on this project we also aimed at developing a methodology to improve the performance of frame-based configuration expert systems. We achieve our goal by knowledge compilation, a computer-aided software engineering (CASE) technique described in Lirov (1991).…”

“…However, this architecture causes major run-time performance degradation, because of the inherent inefficiency of the process implementation (Louis & Vauclair 1988). Therefore, we apply expert system reification (Lirov 1991), a knowledge compilation mechanism that exploits the generality of expert system architecture while upgrading run-time performance.…”

“…In this paper we restrict ourselves to the automatic development of a more efficient Prolog code for a frame-based knowledge representation, although in Lirov (1991) we have also shown how to translate a rule-based reasoning procedure into a low-level language compilable code using knowledge about the target language, namely C. The expert system reifier consists of two modules: a partial evaluator and a compiler (Figure 3). The objective of partial evaluation is to reduce the meta-level overhead by performing as much computation as possible beforehand, independent of the input.…”