KNOWLEDGE AND SKILLS IN DESIGN AND PRODUCTION Emmanuel CAILLAUD and Lionel FRANCHINI Ecole des Mines d'Albi-Carmaux Campus Jarlard, route de Teillet 81013 Albi CT Cedex 09 Phone: (33) 5 63 49 30 92 Fax: (33) 5 63 49 30 99 E-mail: {caillaud, franchin}@enstimac.fr Abstract: Knowledge and skills are considered as important by the firms. In this paper, we present a synthesis of our works on knowledge engineering and skill management. General methodologies are presented. Applications on knowledge engineering for design and skill management in production illustrate the proposed methodologies. Keywords: Knowledge, Skill, Production, Design, Decision Support System. 1. INTRODUCTION: DATA, INFORMATION, KNOWLEDGE AND SKILLS Many works underline the importance of knowledge and skills the firms. Several French firms such as CEA, Renault SA, SNR, Aérospatiale, PSA developped their own approach for knowledge engineering (Ermine, 1996; Fouet, 1997) and other firms such as Sollac developed original approaches for skill management. This importance of human resources is also taken into account in other firms such as BMW, British Aerospace and Nike in many different countries. Moreover, the development of fusions, partnerships and concurrent engineering lead the firms to the problems of knowledge sharing. It is important to distinguish the different terms used: data, information, knowledge and skills. After Franck (1999), we consider data as the basis of knowledge. Comparing, combining and evaluating data transform them into information. Information can be considered as the elementary piece which can be transformed into knowledge. Knowledge can be considered as a set of information structured in order to solve a problem. After Bruneau and Pujos (1992), different types of knowledge can be defined: theorical knowledge, know-how and behaviour. After Le Boterf (1995), skill is characteristic for a human being using several types of knowledge to solve a problem in a given context. In the following parts, we present a synthesis of the different approaches we have developed these last years for knowledge engineering and skills management. In the last part we illustrate our presentation with applications. 2. KNOWLEDGE ENGINEERING Knowledge engineering aims to develop knowledge based systems efficiently. Knowledge engineering methodologies are based on artificial intelligence and cognitive science (Aussenac et al., 1996). Let us consider two of the existing methodologies used in Europe for knowledge extraction which are MACAO and KADS, in order to set up the basic principles. MACAO (Aussenac, 1989) is a general method to support knowledge extraction. The aim of MACAO is to build a conceptual complete model of know- how from the capture of knowledge in various ways (interviews, analysis of data,...). MACAO is applied in four main steps: - extraction and modeling of knowledge from data in the domain,