Author(s)

F. Sandoz, P. Quarrey & N. Aggarwal

Abstract

In this paper, we propose a problem solving method which allows learning by synthetization and case-based reasoning. The synthetization process generates synthetic knowledge which summarizes the solving of a given problem. The synthetic knowledge is then used to solve analogical problems. Case-based reasoning improves the solving process by recovering the steps of solving associated to an analogous problem. The theory of fuzzy logic is used to represent incertain information by linguistic variables and to estimate the analogy between the current problem and a typical problem. This proposition of problem solving method is illustrated through a banking application : PATRILOG, software concer

Keywords