Using fuzzy rules for knowledge representation in oil and gas subject area intelligent systems

Abstract

In the paper it is analyzed how to represent the structure of knowledge about oil and gas subject area in the
form of rules for intelligent systems as well as databases or knowledge bases for information systems. The benefits
of using fuzzy logic rules with linguistic entry, alongside with common approaches such as clear-cut rules, uncertainty
rules, and rules with confidence factors, are reasoned. The formal approach to developing intelligent systems
employing fuzzy sets and inference engine based on the methods and means of fuzzy logic is described. The fuzzification
process of the input crisp data concerning oil and gas related objects is structured, difficulties of the output
fuzzy inference making are defined, compositional operations features are described, choice of defuzzification
method for intelligence system to present realistic and accurate output is proved. Since any knowledge is surely
backed by subjective evaluations, extracting crisp knowledge from oil and gas subject area experts is defined to be
the main challenge facing intelligent systems developing while employing fuzzy approach. Following ways to solve
the investigated issue are offered: developing tools for structured and unstructured (poorly structured) data control
and distribution; clear out sorting data into crisp and fuzzy; providing automated methods for generating linguistic
rules of entry; developing tools for verifying knowledge bases as to their completeness and consistency; using the
membership function verification.