Fuzzy set and possibility theory-based methods in artificial intellingence
Résumé
Fuzzy logic is almost 40 years old and was developed largely outside mainstreamartificial intelligence (AI). As a consequence, fuzzy logic methods are not always well-understood by the AI community, in spite of some valuable efforts such as the organizationof fuzzy logic workshops in AI conferences, e.g., [14,43,50,51]. Yet, an important partof the concern and research in fuzzy logic and possibility theory does focus on issuessuch as knowledge representation, approximate reasoning and reasoning under uncertainty,which are central to AI. These research concerns are indeed less widely known than theengineering applications of fuzzy if-then rules, where control laws are synthesized bymeans of interpolation techniques based on matching degrees [46,59,61]. Originally, fuzzylogic control was very close in spirit to expert, AI-based control. This field has howeverdrifted away from main AI interests in the recent years, and the emphasis on knowledgerepresentation issues in these works now tends to fade. Fuzzy control more and morebecomes a subchapter of nonlinear control.In contrast, this Special Issue is meant to bridge the gap between mainstream AIand current fuzzy set research, and to provide an organized view of recent works bygathering representative applications of fuzzy set and possibility theory-based methodsto AI problems. Before providing a brief presentation of the contents of the special issue,basic facts and key issues in fuzzy set and possibility theories, which are relevant for AIapplications, are first recalled and emphasized.