A Multi-Agent Approach for modeling a Quality Management System with UML Class and Sequence Diagram

In this article, a multi-agent approach modeled by the UML (Unified Modeling Language) is used to solve the piloting problem of Quality Management System QMS. Indeed, this multi-agent modeling of QMS proposes the use a micro framework via a dynamic diagram and macro through a static diagram for the agents development. From this modeling, an information system can be implemented in order to ensure the QMS conformity and effectiveness to the requirements of the series standards ISO 9000, ISO 9001, ISO 9004. In this context, an industrial company example will be proposed and the improvement prospects will be illustrated there. In an increasingly competitive international context,the companies constantly need to adapt and optimize their industrial tools in order to increase their productivity. In particular, the QMS piloting is evaluated according to several parameters (cost, time, customer satisfaction, performance indicators, audit score, quality control, etc....) taking account its contribution to defining the yield and hence the company competitiveness. The need increased for flexibility, agility and effectiveness of QMS results in a growing complexity which it is necessary to know the model and so to control it. In this context, the object oriented modeling is a method which already proved its worth [1]. Indeed, the software and systems designers have a modeling methodologies and languages for the various components system design. For that, the standard modeling UML name was imposed since it is known the most and more used modeling languages [2]. In addition, the Multi-Agent Systems (MAS) designers use another language Agent UML which is not other than an extension of UML language [3] [4]. Indeed, the agent-oriented simulation replaces gradually the various techniques of objects oriented simulation. This is due to several factors that are: (a) its capacity to apprehend many different individual models, from very simple entities to more complex entities; (b) its facility with which various representation levels can be handled by one of the various models. This diversity makes MAS the choice support for the complex systems simulation. Moreover, the UML agent modeling integrates the differences which exist between the agents and objects [5]. We can cite: (a) the MOTIV-PTA application evoked by [6] which uses the UML agent concept to a real application presenting the behavior concepts inside an agent and its relationship with outside; (b) the UML Agent application to the logistics chain management in order to define a methodology and tools to help the designers, to evaluate and to show that UML agent can be applied to the real applications [7]; (c) the multiagents piloting through a holonic approach for a manufacturing system of safety glass [8]. To manage a QMS, a MA approach is used since the correspondence existing between the physical entities (process) and the agents [9]. It is thus possible to represent, analyze and study the MA piloting for a QMS [16]. This paper is organized as follows. Section 2 presents a brief recall on the new multi-agent approach for modeling QMS. Section 3 proposes UML class and sequence diagram. In section 4, we present a real case as an example to validate our model. 2. A brief recall on the new multi-agent approach for modeling QMS Referring to the identified requirements of ISO 9000, ISO 9001 and ISO 9004 [10] [11] [12] and the foundations of the process approach; we can identify five types of processes namely: A direction process that will listen to customer expectations, setting objectives, providing resources, decision-making and confirmation of actions for improvement.