Fuzzy-logic diagnostic rules: a constrained optimisation viewpoint

This paper discusses a knowledge-base encoding methodology for diagnostic tasks. In particular, it transforms fuzzy rules, provided by human experts, into algebraic equalities and inequalities. In this way, the “possible” disorders are the solution of a constraint satisfaction problem. If the disorders are weighted by some a priori possibility assertions (also provided by the experts), then the problem may be cast as a constrained optimisation one. In this way, the need of fuzzy inference systems or “uncertain”-logic schemes is no longer present in the particular setting in this paper. The problem is solved by efficient, widely-known, linear and quadratic programming tools, in principle able to cope with very large-scale problems.