Thus, it is reasonable to focus on promising 500 edges. One simple way is to use an edge set, which is called a 501 ..... In summary, GEAX has four variable parts, i.e., CD, NABL, 734. RSAB, and EC:XS, where NABL is fixed as 100 in this paper. 735. T
Autonomy guarantees that an agent will pursue its goals proactively, so as to accomplish its tasks. Accordingly, the ... where global system properties can be embodied. This suggests that, in the context of ... suffice: in order to design societies,
and this electronic version is released by Control Systems Principles, via its website www.control-systems-principles.co.uk, and ... The potential range of a book on modelling is enormous, and to make the task manageable I have been selective. ... Pr
AbstractâThe multiagent optimization system (MAOS) is a nature-inspired method, which supports cooperative search by the self-organization of a group of compact agents situated in an en- vironment with certain sharing public knowledge. Moreover, ea
Integration Model. Electronic government (E-government) is viewed in this study as the utilization of. Information Communication Technologies (ICTs) to transform the efficiency, .... during fieldwork in domestic command post exercises and uses two ca
Aug 28, 2011 - Learning. Khan, Basit Ahmed. Doctoral Thesis. Submitted for the Partial Fulfilment of the Requirements for the Degree of philosophiae doctor ... C3: Applying the theories related to Place and Experience into practice by extending Multi
cation problems in the smart grid. For example, DCOP so- lutions to electric vehicle charging problems have been pro- posed in [10, 12], and Rust et al. ...... Constraint Programming solver4 as a subroutine. The effect. âp of each device p (see Equ
equilibrium bidding strategy, which is a very questionable assumption (Bajari and ... complex environments that are beyond the reach of lab experiments, does ...
Formative Assessment has been adopted as an alternative to the traditional model based on cumulative exams and has revealed itself as a crucial importance to distance education, by helping the learners' behavior perception and the problems identifica
niques, as proposed by Arnold et al. in  and refined in , use mathematic ...... Chapter 12. Assuming that fast data access is guaranteed, fast arithmetic operations are required to obtain high throughputs. MPC is usually implemented in ......
10#:A2 and 101:A3. These rules define a .... The system uses ts to decide when to invoke the Genetic Algorithm. The Action Set size ...... for each p.d.f. using either a Chi-squared test or Kolmogorov-Smirnov test . The closest match ...... 81.2(
Mar 7, 2003 - As canadian specific highways, climat and laws will be simulated, our simulation results will help us to prove the benefits of ITS for canadian ...
Dec 13, 2017 - Residue Number System (RNS) as the internal num- ber representation across all layer evaluations, ... model of digital hardware source code is available for re-use at https://github.mit.edu/mrhamid/6888_ .... taking the modulus causes
revealing sensitive information that can include full cryptographic keys. Moreover, this compromise is stealthy, meaning that it's not immediately detected by the system owner or defender. We have previously introduced a game-theoretic model for this
properties of real investment markets described by the log-normal, Pareto wealth and Levy stock returns distributions and ..... expected return on risky assets. Included here are such issues as asset allocation, .... field that focuses on the intelli
NPTEL >> Mechanical Engineering >> Modeling and Control of Dynamic electro-Mechanical System. Module 1- Lecture 1. Introduction to Dynamic System Modelling. Dr. Bishakh Bhattacharya. Professor, Department of Mechanical Engineering. IIT Kanpur. Joint
System-Dynamics modelling to improve complex inventory management in a batch-wise plant. Zofia Verwater-Lukszo and Tri Susilowati Christina. Delft University of Technology, Faculty of Technology, Policy and Management. 2600 GA Delft, the Netherlands.
Abstract-- We identify three fundamental issues underlying many problems in .... scaleable architecture that provides communications solutions .... 4 of system and agent architectures. System architecture includes such aspects as multi-agent organiza
ions (i.e., quality values) for the service implementation i of service s over some .... degrading its exposed qualities and then again provides good qualities (or vice .... verge to it. This comes about when no explorer agents are present and the do
Sep 18, 2004 - Faculty of Organization and Informatics, University of Zagreb, Croatia/. Fakultet organizacije i informatike, SveuÄiliÅ¡te u Zagrebu, Zagreb, Hrvatska*. Facultv of Electrical Engineering and Computing, Universitv of Zagreb, Zagreb, Cr
ABSTRACT. The south Indian city of Bangalore provides a challenging yet representative context within which to examine issues of governance of urban social-ecological commons. The city was once famous for its numerous large water bodies, which have w
Abstract—Software Process Assessments is a key factor within organizations to determine their current capability/maturity level and to adopt a Software Process. Improvement initiative. Their chance of success using a standard model is determined by a
2, 143Â±158. Research Article. The GARP modelling system: problems and solutions to automated ... however, as the statistical problems arising in arbitrary spatial data analysis potentially apply to any domain. ..... sampling with replacement provide
May 12, 2010 - systems modelling with scenario planning to support the exploration of uncertainty, identify knowledge gaps that set a (subsequent) research agenda, understand the role and usefulness of historical data, and model research outcomes to
Applying Multi-Agent System Modelling to the Scheduling Problem in a Ceramic Tile Factory Authors: Adriana Giret, Estefanía Argente, Soledad Valero, Pedro Gómez, Vicente Julian Abstract The actual ceramic tile sector needs dynamic production processes to offer its clients on-line programming. Thus, companies can manage real-time response about their services and delivery times of required products, tackling a Mass Customization process in which design and sales activities are done before the production stage. The customer service must cover all activities that can improve the client satisfaction (offers, orders, after-sales services, etc.). Moreover, production tasks scheduling in a ceramic tile factory is a complex problem which requires robust and flexible software applications. Recent advances in Multi-Agent Systems applied to general scheduling problems and industrial applications have demonstrated the advantages of the agent technology in complex distributed problems. In this work we present a Multi-Agent System modelling for a scheduling problem in a ceramic tile factory. Our approach tries to improve the production performance, increase the schedule reliability and keep updated schedules. We propose a multi-agent system in which the constituent agents cooperate to find a feasible schedule taking into account on-line orders, factory layout and capacity, time constraints, anticipated demands and constraints imposed by the master plan. The value of our approach is two fold. On the one hand it is useful for defining the production tasks schedule, while on the other hand it can be suitable for simulating purposes, for example to find out whether a customer order is feasible or to figure out different schedules for a specific production lot. Keywords: Process Design, Production Planning, Multi-Agent Systems Address: Departamento de Sistemas Informáticos y Computación, Polytechnic University of Valencia, C/ Camino de Vera s/n, 46022 Valencia (Spain) [email protected]
Biographical notes Adriana Giret is originally from Villarrica (Paraguay) and receives the BS and E.C.S. degree from the Catholic University of Asuncion in 1998 and 1999, respectively. She is a lecturer and a PhD student at the Computer Science Department of the Polytechnic University of Valencia. Her research interests are multi-agent systems, holonic manufacturing systems, and agent-oriented methodologies. Estefania Argente is originally from Valencia (Spain) and receives the BS and MS degrees in Computing Engineering from the Polytechnic University of Valencia in 2000 and 2003, respectively. She is a lecturer and is doing her PhD at Computer Science Department of Polytechnic University of Valencia. Her current research interests are multi-agent systems, agent design, methodologies and software engineering for multi-agent systems, neural networks and genetic algorithms. Soledad Valero is originally from Valencia (Spain) and receives the BS and MS degrees in Computing Engineering from the Polytechnic University of Valencia in 2000 and 2003, respectively. She is PhD student at Computer Science Department of Polytechnic University of Valencia. Her research interests are multi-agent systems, e-commerce and soft-computing techniques. Pedro Gómez is originally from Bilbao (Spain) and receives the MS degree in Computing Engineering from the Polytechnic University of Valencia in 1992. He is a researcher and is doing his PhD at Research Centre on Production Management and Engineering of Polytechnic University of Valencia. His current research interests are scheduling algorithms and methods and multi-agent systems. Vicente Julian is originally from Valencia (Spain) and receives the BS and MS degrees in Computing Engineering from the Polytechnic University of Valencia in 1992 and 1995, respectively. He is a lecturer and obtains his PhD at Computer Science Department at Polytechnic University of Valencia in 2002. His current research interests are in multi-agent systems, agent design, information retrieval and real-time systems.
1 Introduction The ceramic tile sector is probably one of the major strengths in Spain, and fundamentally in the Valencian Community. Probably due to its high concentration, it has been capable of generating a dynamic of competition that has ended in a continuous improvement of the sector. This progress is finally reflected in an increase of the service, the variety of products and in a production cost decrease. One of the main problems of this kind of companies is production programming. Typically, this problem has been modelled trying to simplify to maximum the environment conditions. Nevertheless the related environment is in fact very dynamic and it reflects the dynamic conditions and constant changes of the Ceramic Tile Sector, such as new client requirements, dynamical work entrance, the availability of machines due to breakdown, etc. The multi-agent system model seems to be a suitable framework for dealing with the design and development of an application which is flexible, adaptable to the environment, versatile and robust enough for the efficient management of a production process. The employment of the agent/multi-agent system paradigm has currently increased as an important field of research within the Artificial Intelligence area. Recently, the application of these techniques seems appropriate for solving complex problems which require intelligence. In particular, the manufacturing industry is one of the domains where the multi-agent system technology provides a natural way to solve problems that are inherently distributed. Moreover, problems of this kind are much related within a holonic perspective. Holons and agents are very similar concepts (Giret, 2004). The Holonic Manufacturing approach models the manufacturing system as a composition of whole/part entities called holons (HMS, 1994). A holon is an autonomous and cooperative building block of a manufacturing system for transforming, transporting, storing and/or validating information and physical objects. The holon consists of an information processing part and often a physical processing part. A holon can be part of another holon and it is an autonomous and cooperative entity. In this work we use agent and holon as similar modelling notions. Manufacturing requirements impose important properties on modelling manufacturing systems (HMS, 1994). These properties define functional attributes and specific requirements for the system structure and the system development process which must be considered by the methodology. Bearing these requirements in mind we have studied software engineering methodologies which are best suited for problems of this kind. We conducted a study on Object-Oriented Methodologies, Enterprise Modelling Language and Methodologies and
Agent-Oriented Methodologies. This study has demonstrated that MAS methodologies are good candidates to work with (see section 3). In this work we present a modelling experience using INGENIAS methodology (Pavon & Gomez, 2003) to develop an agent-oriented solution to the scheduling problem in a ceramic tile factory. INGENIAS is a complete MAS methodology that has good performance in the development of complex systems. The rest of the paper is structured as follows: a complete description of the problem to be handled is done in section 2. Then, the multi-agent system approach is described in section 3. Finally, related works and conclusions are detailed. 2
Currently, the ceramic tile sector needs dynamic production processes that allow the selling process to realize an on-line programming (ASCER, 2003). Therefore, companies can offer their clients a real-time response about their services and delivery times of the required products. Thus, a Mass Customization process is tackled as design and sales activities must be taken before the production stage. Most existing Extended Enterprises have initially focused on reinforcing the links and flows between companies that are involved in the same value chain (Macbeth, 1998). This proposal considers the service time needed to satisfy an order in a supply chain as a key factor, because it has been observed that this time has a relevant weight in decision taking processes. Moreover, process automation is very important to improve several aspects in each enterprise of the supply chain (Linthicum, 1999). The production management process improvement required to face the on-line order configuration automation problem in a ceramic tile sector is presented in this paper. The production system in ceramic tile enterprises is weakly flexible itself (Andrés, 2001). Nowadays ceramic tile enterprises do not have a production management system capable of taking advantages of its own characteristics. Therefore an agile and reactive production planning and scheduling system is considered a major issue. This paper proposes a framework based on a scheduler that can take advantages of the distribution, singularity and dynamicity in several processes to reach an agile order management. The ceramic tile productive sub-system is represented as a three stage hybrid flow shop with sequence dependency in which three stages can be identified: press and glass lines (first stage), kiln (second stage), and classification and packed lines (third stage) (Andrés, 2001). The problem of scheduling hybrid flow shops in absence of setup times has been considered
by many authors, see (Vignier, Billaut & Proust, 1999) for a survey. Nevertheless, there are few papers on machine setup times (Allahverdi, Gupta & Aldowaisan, 1999; Yang & Liao, 1999), perhaps because the usual application area are flexible manufacturing systems, where setup times are negligible. However, in many real world cases, setup times and events in general should be considered to improve the scheduling process. Taking these events into account, different approaches based on the multiagent paradigm can be found (Shen, Norrie, 1999). Customers
Dynamic Production Planning/Scheduling
Sales Supplier Negotiator
Master Plan Generator
Production Programming Stage Manager
Stage 1 Monitor
Stage 3 Monitor
Stage 2 Monitor
Figure 1. Dynamic Production Planning/Scheduling Platform
In the ceramic tile industry the master plan is normally used as the major input data to generate the production scheduling. Nowadays, this generation process is characterized by: (i) Weak automatization, where a large number of schedules are static and each of them is a simple manual conversion from the master plan; (ii) Not reactive, where scheduling systems do not face the set of several events happening in a period (break down, supplier fault, environmental impacts such as humidity, temperature, etc.); (iii) Not taking advantage of singularities, so schedules are based on global models that do not consider neither the own peculiarity of each stage of the production process nor the differentiated states of each stage; (iv) Not Distributed/No distribution, as schedules are executed on a single, centralized computer; (v) Myopic, as models are based on one simple objective function. In Figure 1, we show a common architecture of a ceramic tile manufacturing system. Each module represents specific functionalities which all together implement the entire manufacturing system. In this paper we focus in the production programming module presenting a multi agent system modelling of its internal structure.
Multi-Agent Modelling Approach
Complex manufacturing systems consist of a number of related subsystems organized in a hierarchical fashion. At any given level, subsystems work together to achieve the functionality of their parent system. Each component can be thought as achieving one or more objectives. Thus, entities should have their own thread of control (i.e., they should be active), and they should have control over their own actions (i.e., they should be autonomous). Given this fact, it is apparent that the natural way to modularize a complex system is in terms of multiple autonomous components that act and interact in flexible ways to achieve their objectives. Therefore, the agent-oriented approach is simply the best fit. Regarding the ceramic tile manufacturing system, a multi-agent system (MAS) could be used to achieve integrated optimization of the dynamic production scheduling in a ceramic tile floor. Some advantages of MAS are: (i) it enables using different models and methods to solve the scheduling problem in every stage of the manufacturing process; (ii) it may integrate and optimise a range of scheduling objectives related to different processes and it can adapt to changes in the environment while still achieving overall system goals; (iii) it provides a foundation to create an architecture that helps reaching the complexity reduction, flexibility, scalability and fault tolerance needed; (iv) it improves reactivity to events and enables dynamic scheduling problem resolution; (v) it allows rapid response to new system requirements through the addition of new modules or reconfiguration of existing ones; (vi) it enables to dynamically integrate new agents, remove existing ones or upgrade agents; (vii) agents operate asynchronously and concurrently, which results in computational efficiency. In next sections we present the agent-oriented models and methodology we have used in the development process of the scheduling problem of production tasks in a ceramic tile factory. 3.1
Description of the modelling process
Agents are a powerful abstraction tool for the design and construction of a complex system, because they offer an appropriate way to consider systems with multiple distinct and independent components. In this work we present a modelling experience using INGENIAS methodology (Pavon & Gomez, 2003) to develop an agent-oriented solution to the scheduling problem in a ceramic tile factory. INGENIAS methodology is based in MESSAGE (Caire, Coulier, Garijo, Gomez, Pavon, Leal et al., 2002) and employs several meta-models and a meta-model language for constructing models. A meta-model defines the primitives and syntactic and semantic properties to be used in a model. All meta-models are based on
objects, attributes and relationships. INGENIAS methodology also integrates its meta-models in the Rational Unified Process (RUP) for developing software systems and offers a graphical development tool (Ingenias Development Kit, IDK). During analysis and design phases, five different meta-models are used: (i) organization metamodel, that defines how agents are grouped and which are the system functionality and the existing constraints in agents behaviour; (ii) agent meta-model, which describes the particular agents to be used and their internal mental states; (iii) interaction meta-model, that details how agents are coordinated and interact between them; (iv) environment meta-model, that defines what type of resources and applications are used by the system; and (v) tasks and objectives meta-model, that relates the mental state of each agent with its tasks. In this paper we focus in the analysis phase of the production programming (Figure 1) in order to develop a multi-agent system for the production programming process of a ceramic tile factory. In the following subsections we will show analysis diagrams of a distributed, flexible and autonomous production programming system. This system could be easily connected with the other subsystems of the ceramic tile factory in order to implement the agile manufacturing enterprise. 3.1.1
Use Case Diagrams
A use-case diagram provides a snapshot model of a set of system behaviour that meets a user goal. Thus, this description represents a functional requirement, showing what happens, but not how it is achieved by the system. As mentioned above, our study is focused on the scheduling system where four main use-cases can be identified (Figure 2). In the Schedule Creation use-case, a feasible schedule to be carried out in the following weeks is created. This schedule is developed based on the manufacture lots defined in the master plan. Regarding the Schedule Modification use-case, previous schedules that have arisen problems during their execution are modified. Therefore, those schedules are reconfigured in order to adjust them to factory changes. Concerning the Schedule Execution Monitoring use-case, the current weekly schedule in execution is supervised, informing about the arisen problems. Finally, in the Master Plan Alteration Detection use-case, problems that might alter the master plan are detected.
Figure 2. Uses Cases for the Schedule and Control of Production Tasks
The organization model is defined by the organizational goals and tasks; the workflows that determine associations among tasks and general information about their execution; groups, which may contain agents, roles, resources or applications; and social relationships. Regarding the organization model for the Ceramic Tile Factory (Figure 3), the factory organization has been decomposed in several groups focused in the different activities of the company: Design, Commercial, Production and Purchases & Supplies. Each group contains other groups or different roles. For example, the Production group encloses two groups: Production Programming and Production Plant. Besides, the Production Plant contains four roles: Plant Manager, Press Manager, Kiln Manager and Classification manager, where the first one has authority over the rest. In addition, seven workflows are recognized: (i) Ceramic Tile Design, where the product features to manufacture and market during the season are specified; (ii) Sales, where commercial representatives sell factory products and manage orders from customers; (iii) Analysis of Forecasted Sales, in which future demand predictions are obtained from historical sales data, orders, etc.; (iv) Master Plan Definition, where production orders are defined, sequencing the different product lots to be produced; (v) Schedule and Control of Production Tasks, that includes activities such as determining start time and resources allocation for a specific lot production; (vi) Ceramic Tile Production, that covers all tasks related with the production of the different product lots; (vii) Ceramic Tile Storage, where the final products are stored inside warehouses.
Figure 3. Ceramic Tile Factory Organization Model
Figure 4. Scheduling Process Organization Model
In the organization model for the scheduling process (Figure 4), several roles are distinguished: (i) Manager, responsible for the agent organization, it maintains integrity between all agents in charge of defining and controlling the schedule and regulates the cooperation among the different roles; (ii) Production Plant Manager, that maintains
information about actual plant configuration and knows all restrictions and features of each machine and plant element; (iii) Scheduler, that has the ability to schedule tasks and resources; (iv) Schedule Execution Monitor, that supervises actual execution of a schedule in a specific plant; (v) Master Plan Monitor, that controls possible changes in the Master Plan (according to schedule execution, modification and creation errors) and informs the Manager role when it identifies an alteration that must be propagated to the Master Plan Generator Process; (vi) Schedule Modification Controller, that maintains information about changes needed for adjusting the schedule because of failures in the manufacture process; (vii) Lot Planner, that manages all information about the task sequence; (viii) Schedule Creation Controller, that oversees the information about a new schedule order, more specifically about resource assignment for a specific Master Plan Lot.
3.1.3 Interaction Model Interaction model consists of identifying, for each use case, interaction goals, its members (initiator and collaborators), nature and specification (by means of collaboration diagrams). In Figure 5, only the interaction model for Schedule Creation use case is shown, because of lack of space. Initially, the role Manager establishes an interaction process (called ScheduleInitialization, Figure 6a) with the Schedule Creation Controller, asking for the creation of a new schedule. Then, the Schedule Creation Controller informs the Master Plan role that a new schedule is being created, by means of the ExecutingScheduleCreation message. Next, it communicates with the Lot Planner (using the GetTasksSequence interaction, Figure 6b) in order to obtain the sequence of tasks needed to produce the specific lot to be scheduled. Then, the Schedule Creation Controller communicates the Scheduler the deadlines and tasks to be scheduled (SequenceRequest interaction). The Scheduler has to interact with the Production Plant Manager (Machine/ResourceQuery interaction) in order to assign tasks to resources. Once all tasks have been assigned, the Scheduler sends a NewSequence message to the Schedule Creation Controller, indicating that the new schedule has been created. Finally, the Schedule Creation Controller provides the Manager with this new schedule (Schedule message).
Figure 5. Schedule Creation Interaction model
Figure 6. a) ScheduleInitialization interaction; b) GetTasksSequence interaction
Figure 7. a) Schedule Creation Controller Agent Model; b) Lot Planner Agent Model
Regarding agent models, a specific agent has been assigned to each role identified in the organization model. For each agent, its goals, tasks and mental states have to be associated. Figures 7a and 7b show Schedule Creation Controller and Lot Planner agent models. The Schedule Creation Controller agent is in charge of generating a new lot schedule, so it has to
initialize the schedule and then build a final proposal. The Lot Planner agent provides the task sequence for a specific lot. 3.1.5
This model attempts to answer the questions of why, who and how throughout the analysis process: why refers to the goals that are defined for the system; who refers to the agents which are responsible for the goal fulfilment; and how refers to the set of tasks which are defined to achieve the goals. Regarding our analysis, the decomposition of the Schedule Creation Workflow is shown in Figure 8 as an example of the tasks/goals model. The tasks associated are the different steps of this workflow, and are done by concrete roles.
Figure 8. Schedule Creation Workflow
Figure 9. Transitions considered in the Schedule Creation Workflow
Moreover, in Figure 9 the specification of the transitions considered in the above commented workflow is shown. The process starts with the Identify lots task which creates the Generate Lot Schedule goal. This goal will be satisfied at the end of the process with the execution of the Build Final Schedule task. The rest of the tasks are: Initialize schedule which initiates the control of the schedule creation; Get task sequence which allows to obtain the sequence related to a specific lot; Get plant state in charge of obtaining the current state of the plant; and Allocate tasks which provides a schedule proposal according to the retrieved information. 3.1.6
The environment model of the Production Programming organization is shown in Figure 10. The ExecutedScheduleDB internal application is managed by the Production Programming organization in order to store and update executed schedules created by the organization. The ProductModelDB is an external application managed by the Design Group to store the product definition specification in terms of production tasks. The LotPlanner agent uses operations of the ProductModelDB to determine the task sequence needed to produce a given product. The ProductDesignDB is an external application managed by the Design Group to store the product definition specification in terms of materials and design patterns. The Scheduler agent uses operations of the ProductDesignDB to query the bill of materials for a lot of a given product. The SuppliesDB is an external application maintained by the WareHouse Group to manage the warehouse of raw materials. The Scheduler agent uses operations of the SuppliesDB to query the availability of raw materials needed to produce a lot of a given product in a specific time. The MasterPlanDB is an external application managed by the Production Manager. It is perceived by the Manager agent in order to figure out when the ProductionProgramming Group has to initiate a new schedule creation process. The PlantStateDB is an external application managed by the Production Plant Group to maintain the plant status updated. The PlantManager agent uses operations from the PlantStateDB in order to figure out whether a schedule modification is needed.
Figure 10. Environment model
In the last ten years, an increasing amount of research has been devoted to holonic and agent based manufacturing over a broad range of both theoretical issues and industrial applications. We can divide these research efforts into two groups (McFarlane & Bussman, 2003): Control Architectures and Control Algorithms. Some examples of control architectures are: PROSA (Van Brussel, Wyns, Valckenaers, Bongaerts & Peeters, 1998), the agent based architecture of Bussmann (Bussmann, 1998), agents and function blocks of Deen and Fletcher (Fletcher & Deen, 2001), MetaMorph (Maturana & Norrie, 1997), INTERRAP based architecture (Fischer, 1998). The developments about Control Algorithms range over (McFarlane & Bussman, 2003): Planning and Scheduling (Gou, Hasegawa, Luh, Tamura & Oblak, 1994; Biswas, Sugato & Saad, 1995), Execution and Shop Floor Control (Gayed, Jarvis & Jarvis, 1998; Heikkila, Jarviluoma, & Hasemann, 1995), and Machine and Device Control (Tanaya, Detand & Kruth, 1997; Tanaya, Detand, Kruth and Valckenaers). In spite of the large number of developments reported in these areas, there is very little work reported on modelling manufacturing systems with a Software Engineering Methodology, although its benefits. To date, many of the developments in the manufacturing field have been conducted in an almost “empirical way”, without any design methodology. 5
This paper presents a methodological development, based on multi-agent technology, of an application for the production programming problem in a ceramic tile industry. We have explained that the actual ceramic tile sector needs dynamic production processes that allow the selling process to realize an on-line programming. Our approach is based on a medium time project. Its main goal consists of reaching an order management system in ceramic tile industry able to satisfy each client requirement, adapting the service according to the real state of the distribution/supply chain. This proposal is based on the system capability to offer the most suitable product alternative even though it could involve scheduling changes, if the global quality is improved. In order to satisfy that goal, it is necessary to integrate the different distributed production steps in a flexible, adaptable, versatile, robust and natural way. Agent/Multi-agent systems (MAS) technology has been used in the resolution of this problem, since it provides the required characteristics for manufacturing systems. Specifically, we have presented a
modelling experience using the IDK toolkit of the INGENIAS methodology, which has been successfully employed in other domains. Currently, we have centred our analysis in the programming problem due to its critical importance in the whole process. Nevertheless, as future works we want to analyse the rest of the identified subsystems in a ceramic tile industry. Acknowledgments This work has been partially funded by Polytechnic University of Valencia under grant PIIUPV 5574. References Allahverdi A., Gupta J. & Aldowaisan F. (1999).A review of scheduling research involving setup considerations, Omega, 27, 219-239. Andrés, C. (2001). Production Scheduling in Hybrid Flow Shop with Sequence Dependent Setup Times. Models, Methods and Algorithms. A Ceramic Tile Enterprise Application. PhD Dissertation, UPV. ASCER. (2003). Los sectores español y mundial de fabricantes de baldosas cerámicas 2003. Asociación Española de Fabricantes de Azulejos y Pavimentos Cerámicos. Biswas, G., Sugato, B., & Saad, A. (1995) Holonic Planning and Scheduling for Assembly Tasks. TR CIS-95-01, Center for Intelligent Systems, Vanderbilt University. Bussmann, S. (1998). An Agent-Oriented Architecture for Holonic Manufacturing Control, Proc. of 1st Int. Workshop on Intelligent Manufacturing Systems, EPFL, 1-12. Caire, G., Coulier, W., Garijo, F., Gomez, J., Pavon, J., & Leal, F. et al. (2002). Agentoriented analysis using message/uml. Proc. of the 2nd International Workshop on AgentOriented Software Engineering. Lecture Notes in Computer Science, 2222, 119-125. Fischer, K. (1998). An Agent-Based Approach to Holonic Manufacturing Systems. L. M. Camarinha-Matos, H. Afsarmanesh and v. Marik (Eds.) Intelligent Systems for Manufacturing. Multi-Agent Systems and Virtual Organisations, 3-12. Fletcher, M., & Deen, M.S.(2001). Fault-tolerant holonic manufacturing systems. Concurrency and Computation: Practice and Experience, 13, 43-70. Gayed, N., Jarvis, D., & Jarvis, J. (1998) A Strategy for the Migration of Existing Manufacturing Systems to Holonic Systems. Proceedings of IEEE International Conference on Systems, Man and Cybernetics. 319-324.
Giret, A., & Botti, V. (2004). Holons and Agents. Journal of Intelligent Manufacturing, 15, 645–659. Gou, L., Hasegawa, T., Luh, P., Tamura, S., & Oblak, J. (1994) Holonic Planning and Scheduling for a Robotic Assembly Testbed. Proc. of the 4th Rensselaer International Conference on Computer Integrated Manufacturing and Automation Technology. Heikkila, T., Jarviluoma, M., & Hasemann, J. (1995). Holonic Control of a Manufacturing Robot Cell. Technical report, VTT Automation. HMS, P.R. (1994). HMS Requirements, http://hms.ifw.uni-hannover.de/ Linthicum D.S. (1999). Enterprise Application Integration, Addison-Wesley Information Technologies Series, ISBN 0-201-61583-5. Macbeth. (1998). Partnering Strategy Implementation in the Supply Chain, Strategic Management of the Manufacturing Value Chain, Bittici U., Carrie A. (Edits.) Kluwer. Maturana, F., & Norrie, D. (1997). Distributed decision-making using the contract net within a mediator architecture, Decision Support Systems, 20, 53-64. McFarlane, D., & Bussmann, S.(2003). Holonic Manufacturing Control: Rationales, Developments and Open Issues. Agent-Based Manufacturing. Advances in the Holonic Approach. Springer-Verlag, 301-326. Pavon, J., & Gomez, J. (2003). Agent Oriented Software Engineering with INGENIAS. Multi-Agent Systems and Applications II, Lecture Notes in Artificial Intelligence, 2691, 394–403. Shen W. & Norrie D. H. (1999) Agent-Based Systems for Intelligent Manufacturing: A State of the Art Survey, Int. Journal of Knowledge an Information Systems, 1(2), 129-156. Tanaya, P., Detand, J., & Kruth, J. (1997). Holonic Machine Controller: A Study and Implementation of Holonic Behaviour to Current NC Controller. Computers in Industry, 33, 325-333. Tanaya, P., Detand, J., Kruth, J., & Valckenaers, P. (1998) Object-Oriented Execution Model For a Machine Controller Holon. European Journal of Control, 4, 345-361. Van Brussel, H., Wyns, J., Valckenaers, P., Bongaerts, L., & Peeters, P. (1998). Reference Architecture for Holonic Manufacturing Systems: PROSA. Computers In Industry, 37 255-274. Vignier A, Billaut JC & Proust C. (1999). Les Problemes d´ordennancement de type flow shop hybride: Etat de l´art, RAIRO Operations Research, 33, 117-183. Yang W. & Liao C. (1999). Survey of scheduling research involving setup times, International Journal o Systems Science, 30, 2, 143-155.