بهبود مستمر از طریق تجزیه و تحلیل دانش هدایت شونده در بازخورد تجربه
|کد مقاله||سال انتشار||تعداد صفحات مقاله انگلیسی||ترجمه فارسی|
|6836||2011||13 صفحه PDF||سفارش دهید|
نسخه انگلیسی مقاله همین الان قابل دانلود است.
هزینه ترجمه مقاله بر اساس تعداد کلمات مقاله انگلیسی محاسبه می شود.
این مقاله تقریباً شامل 8860 کلمه می باشد.
هزینه ترجمه مقاله توسط مترجمان با تجربه، طبق جدول زیر محاسبه می شود:
Publisher : Elsevier - Science Direct (الزویر - ساینس دایرکت)
Journal : Engineering Applications of Artificial Intelligence, Volume 24, Issue 8, December 2011, Pages 1419–1431
Continuous improvement in industrial processes is increasingly a key element of competitiveness for industrial systems. The management of experience feedback in this framework is designed to build, analyze and facilitate the knowledge sharing among problem solving practitioners of an organization in order to improve processes and products achievement. During Problem Solving Processes, the intellectual investment of experts is often considerable and the opportunities for expert knowledge exploitation are numerous: decision making, problem solving under uncertainty, and expert configuration. In this paper, our contribution relates to the structuring of a cognitive experience feedback framework, which allows a flexible exploitation of expert knowledge during Problem Solving Processes and a reuse such collected experience. To that purpose, the proposed approach uses the general principles of root cause analysis for identifying the root causes of problems or events, the conceptual graphs formalism for the semantic conceptualization of the domain vocabulary and the Transferable Belief Model for the fusion of information from different sources. The underlying formal reasoning mechanisms (logic-based semantics) in conceptual graphs enable intelligent information retrieval for the effective exploitation of lessons learned from past projects. An example will illustrate the application of the proposed approach of experience feedback processes formalization in the transport industry sector.
Industrial products developed nowadays are more and more complex and involve several technologies at the same time. Moreover, design time is reduced, adding new constraints during pre-industrialization phases. In this context, sharing experience feedback and lessons learned is a key issue to improve the performance of organizations over time. However, sharing this knowledge is made difficult in large organizations for two main reasons: • the project based management which creates a partitioning of the produced knowledge, • the distributed structure of nowadays organizations implies virtually space across geographic and temporal boundaries. In order to overcome these difficulties, building an experience feedback and lessons learned repository can be of major interest to share knowledge through time and space. This is made all the more relevant that, during the past decades, considerable efforts have been made by industrial firms in order to standardize their products and their processes. Therefore, from a representational point of view, the knowledge acquired from previous problem solving experiences should be reused as much as possible to allow the domain experts to find appropriate solutions with minimal effort. After solving one problem (leading to an experience) of many to be solved, experts can transfer lessons learned from one context to another without having to achieve the whole problem solving process. However, in some fuzzy domains, experts may sometimes be more overconfident and they may miss very obvious features without a root cause problem analysis or with a misleading problem analysis. These new constraints are rarely taken into account in traditional problem solving methods. The concern of this work is to address the knowledge capitalization and exploitation for continuous improvement in the resolution of industrial problems. Different tools and approaches for the acquisition, representation and exploitation of knowledge have been proposed especially in knowledge engineering sciences (Hicks, 2004). However, these methods dedicated to model expert knowledge modeling, show some practical difficulties: experts often lack motivation, skills and time to document their expertise, a mediator is often needed to remove semantic distance between the expert and the knowledge-based system, the regular update of the knowledge referential is difficult. Thus, experience feedback, which advocates a capitalization during the activities of experts, helps to overcome these disadvantages (Henninger, 2003). Naturally, the captured knowledge remains fragmentary and requires additional efforts if it is to be generalized. Finally, a compromise appears between the quality and generality of knowledge and the effort required to acquire it. In a context of rapidly evolving knowledge (such as encountered in continuous improvement processes), it may be interesting to focus on reducing the effort to obtain knowledge allowing experience feedback (Weber et al., 2001). Besides, in many companies, quality certification requirements have led to standardized problem-solving processes in which experts investigate the causes of the problems and attempt to eradicate them. In this context, the experience feedback approaches based on standardized problem solving methods can contribute to continuous improvement in business processes. In an experience feedback approach of this kind, the knowledge is generated, on one hand, from the capitalization of knowledge and know-how used in industrial processes and, on the other hand, formalized through the tools and methods used by actors in their work (Jacobsson et al., 2010). For example, in the Swedish Center for Lessons Learned from Incidents and Accidents (NCO), learning from accidents is institutionalized in order to overcome various social barriers and to disseminate information so that new insights in accident prevention are as widely applied as possible (Lindberg et al., 2010). Historically, experience feedback was mainly based on statistical methods to identify some failure laws. However, this kind of feedback does not allow the extraction of expert knowledge from the technical data. This is made possible by the “cognitive approach” of experience feedback modeling. It models the expert knowledge of the organization and facilitates the enrichment of knowledge repository by using methods from artificial intelligence. The cognitive vision framework of experience feedback provides means of understanding, interpreting, storing and indexing the activities of experts (Weber and Aha, 2003). This work specifically focuses on issues in the “analysis” activity (mainly oriented towards the search of the root causes of a problem) of experience feedback processes. It uses semantic technologies and reasoning mechanisms to refine indexation and adaptation steps by keeping track of the analysis performed. The analysis model must incorporate the possibility for an expert to appoint the most significant descriptors necessary for the best explanation of factors affecting problem occurrence and severity (Beler, 2008). The resulting analysis would correspond to a combination of relevant pieces of cognitive task analysis on which the domain expertise has associated a degree of belief that takes into account all the available evidence (Shafer, 1990). Indeed, knowledge related to cognitive elements underlying the analysis generation and lessons learned can be produced by tools that enable the formal description of physical tasks and cognitive plans required from a user to accomplish a particular work goal (Militello and Hutton, 1998). The paper is structured as follows. Section 2 exposes a state of the art concerning knowledge management for experience feedback and a comparison between the potentially relevant semantic technologies is discussed. Section 3 presents the three-layer model proposed for analysis improvement in experience feedback framework. An illustrative application example is exposed in Section 4. Finally, Section 5 concludes and discusses future challenges
نتیجه گیری انگلیسی
As presented in this paper, the proposed approach relates to knowledge management in problem solving and experience feedback processes. The main contribution is the proposition of a dedicated analysis model relying on three complementary layers: operational, semantic and belief. This model helps to support experts: • When they look for the most plausible root causes for a problem (use previous analysis),when they elaborate an action plan to solve and eradicate the problem (use previous solutions). • More specifically, with this cognitive approach of experience feedback, root causes are reasonably identifiable and the analysis model associated to lessons learned indicates the arguments in favor of the chosen solution according to experiential knowledge exploitation. There are several practical industrial benefits of the proposed technologies/methodologies experimented in the railway industry sector: • The description of the basic elements of root cause at the semantic level prevents a model analysis to be misunderstood and facilitates the reasoning processes over the expert opinions. Moreover, the semantic enhancement of experience feedback modeling makes it possible to better identify major issues for industrial processes improvement. • The knowledge exploitation is possible asynchronously and remotely throughout the organization: the expertise assets are enriched over time and shared. • The explicit integration of the expert opinions on the plausibility of hypotheses during root cause analysis enables to better take into account this expertise for reasoning. The use of the Transferable Belief Model and related fusion mechanisms facilitates the inference. However, some work still remains to make this representation more easily accessible to practitioners. From a more global perspective, the knowledge engineering technology implemented enables to collect and exploit experiential knowledge in continuous improvement processes of any complex system in which problematic events require in depth (expert) analysis. Several issues requiring additional efforts are currently under investigation: • an improved support of experts when they split masses between hypotheses since the use of the TBM at an operational level may not be intuitive, • the active dissemination (push) of experiences to the relevant actors by integrating the actor profiles (expertise domain, competences), • the coupling of the experience feedback model with specific architectural principles (Krishnan and Bhatia, 2009) to foster better interoperability with business applications.