Towards a Dynamic Reference Framework for Product Information Systems: Focus on Engineering Change Process
نویسندگان
چکیده
This article deals with a methodological framework for Product Information Systems (PIS) Engineering. The objective is to propose an approach for developing PIS by reuse of "patterns". These patterns form generic solutions to problems frequently occurring during PIS specification and implementation. Their use would contribute to the rapid development of numerous PIS. A special interest is given to identify and specify different business patterns for reuse at the time of specification. The aim is to identify them from the field's activity. The first step in the process of patterns engineering consisted of constructing a reference framework, specifying the various PIS concepts. This framework is based on a static framework providing a common terminology and modelling the principal concepts used to describe products and related processes and on a dynamic framework modelling the behaviour of product concepts in response to processes execution. It forms a basis for exploring the most recurrent problems during PIS specification in order to specify the associated patterns. In this paper, we focus on the dynamic framework and we illustrate our approach with a particular business pattern developed to specify the versioning dynamic resulting from engineering change process. Résumé: Nous présentons dans cet article un cadre méthodologique pour l'ingénierie des Systèmes d'Information Produit (SIP) basé sur la réutilisation de "patrons". Les patrons constituent des solutions génériques à des problèmes fréquemment rencontrés lors de la spécification et l'implémentation des SIP. Leur utilisation devrait contribuer au développement rapide de plusieurs SIP. Un intérêt particulier est porté à l'identification et la spécification de différents patrons métiers pour la réutilisation lors de la spécification. L'objectif est de les identifier à partir de l'analyse du domaine. La première étape dans ce processus d'ingénierie des patrons a consisté à construire un référentiel du domaine spécifiant les divers concepts gérés dans les SIP. Ce référentiel est basé sur un référentiel statique fixant la terminologie et modélisant les principaux concepts utilisés pour décrire les produits et les processus associés et sur un référentiel dynamique modélisant le comportement de ces concepts. Ce référentiel constitue une base pour explorer les problèmes fréquemment rencontrés lors de la spécification des SIP afin de proposer les patrons associés. Dans ce papier, nous nous focalisons sur le référentiel dynamique et nous illustrons notre approche par un patron métier dédié à la spécification de la dynamique de versionnement résultant du processus de modification de produit. 1. Context Industrial companies must today obtain a rigorous control of the whole product related information in order to increase their reactivity to the different changes during the product life cycle. This context makes Product Information Systems (PIS), developed to manage product engineering data, a strategic stake for industrial organizations. Since the mid 1980’s, a class of software applications, called Product Data Management Systems (PDMs) has emerged. PDMs are tool platforms adapted to characterization of items, Bills Of Materials, documents, procedures, etc., enabling thus implementation of PIS. 1.1. PIS engineering Problems: Several research works have been already interested in management of product engineering data. These contributions are particularly concerned with problems of integration/sharing data [1,2] and PDMs deployment [3], problems that are extremely sensitive in PIS development. However, few are the contributions that deal with the actual engineering of the PIS, i.e. the methodological aspects relating to the design and implementation of such systems. The development of such systems implies significant investments for the companies. The stakes associated with these methodological aspects are therefore strategic also. Indeed, PIS specification and implementation are not tackled in a methodical way of computerization; they raise many difficulties: lack of formal specification models for users requirements formulation, lack of approaches able to draw up clear and unambiguous specifications, absence of continuity in specification when evolving from analysis stage (functional specifications) to design stage (technical specifications), complexity in PDMs implementation, slowness in development, etc. Furthermore, we should highlight that the current economic competition context tends to reduce product life-cycles more and more. Industrial companies are thus obliged to develop numerous PIS applications within reduced times, which will remain operational only during the lifetime of the product(s) concerned. This research is thus focusing on the definition of a methodological framework for PIS engineering: offering models of various abstraction levels that cover the complexity of such systems, support the communication between actors and ensure a continuum of transformations during specification, enabling “deviation” development of such systems, i.e. make it possible to design and produce new PIS from PIS already designed and produced, thus allowing re-use of existing software components and specification elements, in order to accelerate the PIS development process. 1.2. A methodological approach based on patterns reuse: An efficient way to speed up PIS development consists of allowing “deviation” specifications, both as regards capitalization and reuse of concepts already encountered and consideration of available software resources (components and systems). Consequently, the reuse approach, already effective in software engineering, is a key factor to our PIS development methodological approach, both for specification and implementation of PIS and for their evolutive maintenance. A general definition of the term reuse could be "a new development approach by which a system can be built from existing components already described, carried out, tested and accepted in past experiences". The aim is to avoid remaking all with each new application or when changing a software. Thus, we deliberately are in the framework of a capitalization and reuse approach of functional, organizational and software components, at all stages of the PIS development cycle. Today the reuse approach is extensively used in software engineering, and different forms of reusable software components have already been proposed. Three major approaches can be identified: toolkits [4], frameworks [5,6] and patterns [7,8]. A special interest is given to the pattern approach. In the sequel, we present the concept of pattern and show its advantages over the other forms of reusable components. Several pattern definitions are presented in the literature [8,9]. It comes out from these definitions that a pattern constitutes a know and know-how base to solve a problem frequently occurring in a particular field. This base allows to identify the problem to solve (for example BOM management), to propose a correct solution to take it into account (for example by specifying software components) and finally to give indications to adapt this solution in a particular context. Patterns form thus "know-how" oriented components, while other forms of reusable components are only "know" oriented components. These latter provide only solutions to problems whereas "know-how" oriented components provide solutions but also the manner to construct these solutions. Patterns form hence an engineering guide by organizing hierarchically and functionally problems and the manner to resolve them. Reuse of patterns is thus, in our opinion, the most suitable form of reuse for PIS engineering, as it can be used in all stages of the PIS development cycle (analysis, design, implementation). The objective is to adapt the pattern approach to a particular field, that is of PIS, according to a target technology, the PDMs. The proposed methodological framework is then based on a consistent set of models of various abstraction levels that will be developed by reuse of patterns. Each level proposed must enable a problem to be solved (of a functional, organizational, technical nature, etc.) specific to PIS development. Two forms of patterns must be taken into consideration in PIS design: • business patterns specifying the organizational information system (OIS) and important in the stages of formulation of needs and definition of functional specifications. They provide solutions for application field problems. They must so be able to take into account a set of information needs associated both with product and various business processes acting on it. Two forms of modeling are thus essential: product modeling and process modeling. • software patterns specifying the part of the OIS which will result in computerization. Definition of software patterns is strongly linked to the PDM system which is at the basis of PIS development. The modeling based on software patterns is the expression of a technical solution that takes two major problems into account: implementation of product and process models and communication of the PIS with other systems. With respect to implementation of models, the PDM systems use database models (relational or object) and workflow models. In both cases the models proposed in tools are extensively used and can thus be integrated in the proposed methodological framework. Definition of software patterns associated with PIS are based on the functions proposed by most PDM systems. With respect to business patterns, the aim is to identify them from the field’s analysis. We give interest in the present research to business patterns. 1.3. Business patterns engineering: Introducing patterns in PIS engineering process requires first to construct a pattern library or catalogue. In this order, a new activity is required and consists to identify and specify these patterns. Once the pattern catalogue is built, this constitutes a formalization for the PIS engineering process. Highlighting Business patterns requires first, identification and comprehension of the problems common to PIS (related to product definition, representations and processing) and second, specification of the associated solutions. The first step, consisting of identifying problems from the PIS field, is based on the study of existing PIS models. There have been many proposals for product [10,11,12] 1 A pattern catalogue is a set of patterns organized with relations of various natures such as use, extension/refinement and alternative relations [13]. and process modeling [14,15,16]. However, we noted that most of these contributions do not directly address the issues arising in Product Information Systems field – the area where we aim to concentrate our efforts. Few are the contributions that discuss the concepts needed in PIS and model the variety of relationships existing between these concepts. We also noted the absence of a consensus on the concepts managed in PIS and their formalization. Then, we focus our efforts on a field analysis, providing a generic framework for the field. This reference framework is based on a static reference framework and a dynamic one. The concepts discussed in the static reference framework are related with product and process descriptions. It proposes first a terminology and semantics of these concepts and then proposes a reference model to fix these concepts. The dynamic reference framework models the behavior of product concepts in response to a process execution. The model thus obtained in the reference framework forms, with the existing PIS models, a basis for exploring the problems frequently occurring in the field. Once identified, these problems are compared to problems treated in existing design patterns catalogues. This allows, on the one hand, to fix the problems already treated in catalogues and, on the other hand, to discover new problems. The second step of the engineering pattern process consists then to specify solutions to the identified problems. This phase is partially based on existing design patterns catalogues. Then, when the identified problem can be brought back into problems raised in existing design pattern catalogues, the corresponding solution is integrated in our pattern catalogue. When a new problem is identified, a new solution is then proposed. 2 We quote mainly the draft standard STEP (ISO 10303) developed by the ISO and the Object Management Group initiative [17] aiming at developing a standard of product data representation allowing exchange and sharing of these data. 3 This reference framework was discussed in a previous comuunication [18]. In the sequel, we focus on the dynamic reference framework. 2. Dynamic Reference Framework Engineering Change Process The dynamic reference framework provides generic models for specification of PIS dynamic when activating various PIS processes. In this section, we focus on the engineering change (EC) process, one of the most important process in PIS field. Our interest is to understand the dynamic resulting from that process and then to model it. The outline of this section is as follows: first, an overview of EC process is given. Then, in a second part, an analysis of the versioning dynamic resulting from that EC process is presented and a pattern is given to specify this dynamic. But above all, the formalism used to model PIS processes and their dynamic is presented. 2.1. PIS processes modeling: We model a PIS process using UML activity charts. In a previous communication [19], a Metamodel fixing a terminology for the concepts used in process modeling is defined. This Metamodel is based on “Activity, Operation and Process” concepts: • An Activity represents any kind of action carried out in the company by one or more resources, with various roles, in order to satisfy a given purpose. An activity can be an operation or a process. By role, we mean the authorisation or right of the resource in the activity in which it takes part and by purpose we mean the result to be provided by the system on which the process acts. • An Operation is an elementary activity which can be carried out separately in the company or gathered with other activities to modify the state of the objects. An operation can not be decomposed because it corresponds to an elementary purpose and it 4 Unified Modeling Language; the standard of modelling in object oriented development [20]. is ensured by a single resource with a single role. • A Process is a partially ordered succession of activities performing a purpose and requiring one or several resources to be completed. A process is a decomposable activity. It is composed of other activities linked by succession relationships which are labelled by conditions of succession. Two kind of succession between activities are distinguished: AND succession: related to an activity which has one or several followings, conditioned by nonexclusive conditions. In case of a single following, it corresponds to a sequence of activities. In case of several followings, it corresponds to an activity whose followings can be done in parallel when the conditions of succession are checked. OR succession: related to an activity with several followings, conditioned by exclusive conditions (disjoined). This case occurs when a decision must be taken: a choice between several alternative activities is made for the continuation of the process, according to the result reached at the former activity. Three natures of OR transitions are distinguished: continuation of the nominal process, redirection or abort of the process of higher level. The above defined concepts are thus modeled in UML activity charts as follows: an activity is represented by an UML activity, stereotyped according to its nature (operation or process). The decomposition of a given process will be represented thus by an activity chart. We suppose that a process has an initial state which marks the start of the process and can have one or more final states corresponding each one to a different condition of process end. A resource is represented by an UML actor, stereotyped according to its role. The transitions between activities are represented by UML transitions stereotyped according to their nature (continuation, redirection, abort). The name, between brackets, corresponds to the condition of transition (for example agreement, refusal...).
منابع مشابه
Developing customer relationship management (CRM) processes using standard frameworks and MCDM techniques
Organizations have found that functional approach to business destroys flexibility and agility by gaining experience over the time. The main weakness with Task-based organizations is that they can hardly act flexibly and adapt themselves to the changing environment. In today’s' dynamic and competitive business world, organizations focus significantly on managing and improving their business pro...
متن کاملSelf-healing in payment switches with a focus on failure detection using State Ma- chine-based approaches
Composition, change and complexity have attracted ev- eryone’s attention towards Self-Adaptive systems. These systems, inspired by the human body, are capable of adapting to changes in the inner and outer environment. The main objective of this study is to achieve a more con- venient availability for e-banking services in the payment switch, using self-healing systems and focusing on the failur...
متن کاملRobust inter and intra-cell layouts design model dealing with stochastic dynamic problems
In this paper, a novel quadratic assignment-based mathematical model is developed for concurrent design of robust inter and intra-cell layouts in dynamic stochastic environments of manufacturing systems. In the proposed model, in addition to considering time value of money, the product demands are presumed to be dependent normally distributed random variables with known expectation, variance, a...
متن کاملSelf-healing in payment switches with a focus on failure detection using State Ma- chine-based approaches
Composition, change and complexity have attracted ev- eryone’s attention towards Self-Adaptive systems. These systems, inspired by the human body, are capable of adapting to changes in the inner and outer environment. The main objective of this study is to achieve a more con- venient availability for e-banking services in the payment switch, using self-healing systems and focusing on the failur...
متن کاملManaging design variety, process variety and engineering change: a case study of two capital good firms
Many capital good firms deliver products that are not strictly one-off, but instead share a certain degree of similarity with other deliveries. In the delivery of the product, they aim to balance stability and variety in their product design and processes. The issue of engineering change plays an important in how they manage to do so. Our aim is to gain more understanding into how capital good ...
متن کاملA social recommender system based on matrix factorization considering dynamics of user preferences
With the expansion of social networks, the use of recommender systems in these networks has attracted considerable attention. Recommender systems have become an important tool for alleviating the information that overload problem of users by providing personalized recommendations to a user who might like based on past preferences or observed behavior about one or various items. In these systems...
متن کامل