EXPERT SYSTEM DESIGN ON THE BASIS OF HIGH-LEVEL KNOWLEDGE ORIENTED MODELS

Константин Владиславович Головань

Abstract


The main points related to the design of the integrated decision supported expert systems are analyzed in the paper. The perspectives of hybrid knowledge representation model are considered. In order to represent the domain knowledge it is proposed to use a high level knowledge oriented model that makes it possible to describe the processes of analysis, mining, and processing of domain knowledge in a form of interaction of some typical predefined functional blocks. The main advantages of the developed functional knowledge-oriented model are: modularity (representation of monitoring, diagnostics and control processes of complex technological systems and objects in a form of separate knowledge-oriented components interaction); universality of the typical functional blocks library (solution of typical tasks, arising in the process of technological object control); adaptability (easy adaptation to a specific domain); openness (gives the user a possibility to set the custom behavior); activity (interaction of typical functional blocks with each other that makes it possible to automate the process of knowledge acquisition and processing and also interaction of functional blocks with a hybrid production-frame model that makes possible to increase the efficiency of knowledge procession during the decision making process). Every typical intelligent element is a functional block with a set of inputs {IN} and outputs {OUT}. The behavior of such block is defined by its purpose. The whole set of typical intelligent blocks that is used in construction of functional knowledge-oriented model according to the block purpose can be divided into several different classes. On the basis of the selected representation model the processes of knowledge formalization are described. The main advantages of the selected approach to formalize the domain knowledge are stated. On the basis of the represented instrumental tool structure the computer the system has been made. The main stages of expert system creation and their key features are considered. The editor of functional knowledge-oriented model has been presented. The basic functions of the editor are model visualization and debugging. Instrumental tool make possible to build control decision expert systems in different domains. The example of expert system in domain of electrochemical protection of pipelines from corrosion has been considered. The basic directions of possible updating of mathematical model and instrumental tools are described in conclusion

Keywords


expert systems; knowledge representation models; hybrid models

References


Chen, C., Song, M. Representing Scientific Knowledge: The Role of Uncertainty. Springer, 2017. 375 p.

Schreiber, G., Akkermans H., Anjewierden A., de Hoog R., Shadbolt N., de Velde W. V, Wielinga B. Knowledge Engineering and Management: The CommonKADS Methodology. MIT Press, 2000. 455 p.

Golovan', K. V. Vysokourovnevye modeli analiza, obrabotki i izvlecheniya znanii v protsesse razrabotki ekspertnykh sistem [High-level models for knowledge analysis, processing and mining in the process of expert system development]. Radioelektronnye i komp'yuternye sistemy – Radioelectronic and computer systems, 2006, no. 13, pp. 46-55.

Prokhorov, A. V., Golovan', K. V. Strukturizatsiya znanii na osnove vysokourovnevykh i produktsionno-freimovykh modelei v ekspertnykh sistemakh prinyatiya upravlencheskikh reshenii [Knowledge structurization on the basis of high-level and production-frame models in control decision-making expert systems]. Systemy obrobky informatsiyi – Information processing systems, 2006, no. 6 (55), pp. 153–157.

Hastie, T., Tibshirani, R., Friedman, J. The Elements of Statistical Learning: Data Mining, Inference, and Prediction. Springer, 2009. 767 p.

Bramer, M. Principles of Data Mining. Springer-Verlag, 2016. 541 p.

Pinaud,‎ B., Guillet,‎ F., Cremilleux, B.,‎ de Runz C. Advances in Knowledge Discovery and Management: Volume 7. Springer, 2017. 147 p.

Ertel, W. Introduction to Artificial Intelligence. Springer, 2017. 370 p.

Kaplan, J. Artificial Intelligence: What Everyone Needs to Know. Oxford University Press, 2016. 192 p.

J. González, E., Acosta Sánchez, L., Hamilton Castro, A. F., Artificial Intelligence Resources in Control and Automation Engineering. Bentham Science Publishers, 2012. 204 p.




DOI: https://doi.org/10.32620/reks.2018.1.08

Refbacks

  • There are currently no refbacks.