Archive TOULOUSE Archive Ouverte ( OATAO )

To take into account the experience feedback on solving complex problems in business is deemed as a way to improve the quality of products and processes. Only a few academic works, however, are concerned with the representation and the instrumentation of experience feedback systems. We propose, in this paper, a model of experiences and mechanisms to use these experiences. More specifically, we wish to encourage the reuse of already performed expert analysis to propose a priori analysis in the solving of a new problem. The proposal is based on a representation in the context of the experience of using a conceptual marker and an explicit representation of the analysis incorporating expert opinions and the fusion of these opinions. The experience feedback models and inference mechanisms are integrated in a commercial support tool for problem solving methodologies. The results obtained to this point have already led to the definition of the role of ‘‘Rex Manager’’ with principles of sustainable management for continuous improvement of industrial processes in companies. * Corresponding author. Tel.: +33 6 24 30 23 37; fax: +33 5 62 44 27 08. E-mail address: [email protected] (B. Kamsu-Foguem). the associated inference mechanisms, necessary to instrument the feedback processes. An experience feedback approach is supported where the relevant knowledge can be shared, categorized and updated for their formal use in knowledge-based engineering applications [9–11]. Most works done in this objective are based on the use of inference mechanisms similar to those proposed in Case-Based Reasoning [12]. This is the case of works proposed by Bergmann [13] and more recently another work described by Armaghan [14]. This point seems very important and we will resume our proposal ideas commonly used in the field of Case-Based Reasoning (similarity search in particular). Finally, the thesis presented in [15] shows the importance of integrating subjective data in order to enrich an information system for the experience feedback. In the same viewpoint, Aven and Zio provide new insights on the treatment of uncertainties in risk analysis context to faithfully represent and express the knowledge available to best support the decision making [16]. In complex domains it is usually quite difficult to introduce contextual information using expert’s rules as background knowledge. However, sometimes that information should be taken into account when making decisions, as it provides some relevant knowledge [37]. Our propositions include both objective and subjective views relating to analysis and the characteristics of analysis, in problem solving processes. In industrial practices, feedback from technical investigations will help improve process management practices and the quality of products or services and, thereby, strengthen feedback and the sharing of experience with the development of a prevention culture, which targets each employee as well as management. In the context of processing various non-conformances, we specified an evaluation phase of the criticality (Table 1) of the problem to be solved to define the means put in place to deal with it. For example, the appropriate analysis of conformances and corrective actions may yield information that could lead to changes in resource examination and maintenance management, thereby improving quality service and the progress of safety programs. As shown in Fig. 1 (below), we consider, in our work, experience that corresponds to a container incorporating the context (problem description and analysis), analysis (expert on finding the cause of the problem) and the solution (set of actions to resolve the problem). Ei and experience will be represented by a triplet Ei = Coi, Ai, Si where Coi, Ai and Si are respectively the context, analysis and solution Ei of the experience. A base of experience corresponds to a set of experiences: BEx p 1⁄4 fEi; i 2 f1; . . . ; ngg. In Fig. 1, we also indicate different inference mechanisms that can be constructed from information stored in the experiments. The first mechanism is an investigation of problems solved in the past using a similarity measure of context. This operation is fairly standard in the reasoning Case-based, but we will propose a particular method, adapted to the subject matter. The second inference mechanism is the intended reuse of analysis expert. On this point, the proposal described in this article has, to our knowledge, never been addressed in the available scientific literature. The third inference mechanism suggested in Fig. 1 is the adaptation of the solutions. This point is not described in the article. The article is structured as follows: in Section 2 we present the proposed context representation (Section 2.1) and the mechanisms associated research previous experience relevant to reuse (Section 2.2). Section 3 is devoted to the description of expert analysis (Section 3.1) and the principles proposed to guide the analysis when a new problem is to be solved (Section 3.2). Finally, a sample application and the software implementation are provided to illustrate these principles (Sections 4 and 5). Section 6 indicates the findings of the study. Table 1 the Criticality Matrix. CRITICALITY= (5*S + 2*R + Q + M)*P S Safety 7 Risk of catastrophic: Fatality or multiple injures 3 Risk of critical: Single fatality, severe injury 1 Marginal: Minor injury