An Intelligent Manufacturing System with Biological ...

International Journal of CAD/CAM Vol. 10, No. 1, pp. 39~50 (2010)

An Intelligent Manufacturing System with Biological Principles Hong-Seok Park* and Ngoc-Hien Tran School of Mechanical and Automotive Engineering, University of Ulsan, Ulsan, Korea Abstract − The next generation of manufacturing systems development is known as intelligent manufacturing systems in which the machines and processes are equipped with the artificial intelligent capabilities. The principles of biology for storing information and implementing cognitive behaviours have been applied to manufacturing systems. This paper presents a novel concept of Intelligent Manufacturing System with Biological Principles (IMS-BP). In order to achieve the IMS-BP, the genetic and intelligent technologies, which are Bio-inspired technologies, such as the swarm intelligence, cognitive agent, machine learning and evolutionary computation are proposed for embedding knowledge in machines, consequently, implementing process planning and maintenance. As a result, the IMS-BP with advanced characteristics such as inheritance, intelligence, adaptation and self-organization is developed in which a novel approach for generating process planning and self-maintenance system is proposed. Keywords: Manufacturing Control, Bio-inspired technologies, Cognition, Evolutionary algorithms, Intelligent Components

1. Introduction The complexity and dynamic of the manufacturing environment are growing due to the changes of product types, suppliers, as well as the unexpected disturbances in the machining or assembly systems. Traditional centralized and sequential manufacturing systems are being found insufficiently flexible to adapt to these challenges [Leitao 2008]. In order to cope with the changes of the manufacturing environment, new methods and technologies have been proposed in which the distributed manufacturing control system and the biological inspired technologies for implementing this system are remarked. Many novel paradigms that are known as intelligent manufacturing systems were proposed in the literature. Genetic [Brussel 1995], Biological [Ueda 2000, Ueda 2007], and Holonic [Leitao 2008, Leitao 2009] manufacturing systems are the remarkable concepts in which the organization structures and mechanisms from the field of biology are transferred into manufacturing environment. In these concepts, the evolutionary algorithms such as Genetic algorithms, Artificial Neural Network, and Ant Colony Optimization were used for implementing the intelligent functionalities such as learning, dynamic scheduling, and optimal problems [Sumichrast 2000, Cus 2008]. However, the current manufacturing systems have the weakness of the centralized control system [Leitao 2008]. The potentials of bio-technologies have not been fully exploited due to their technologies are applied in the separate manufacturing applications, which are either manufacturing scheduling systems or *Corresponding author: Tel: +82-(0)52-259-2294 Fax: +82-(0)52-259-1680 E-mail: [email protected]

manufacturing control systems. In order to make an improvement of the proposed concepts, this paper presents a novel Intelligent Manufacturing System with Biological Principles (IMS-BP) in which the structure of IMS-BP is a swarm of cognitive agents. Consequently, the resources on the shop floor such as machine tools, robots, work-pieces, and so on are controlled by the corresponding cognitive agents. The IMS-BP architecture changes the methods of generating the process planning and maintenance. By this way, the central Computer Aided Process Planning (CAPP) and the central maintenance of conventional manufacturing systems are converted into the distributed process planning and self-maintenance of individual cognitive agents. The IMS-BP is implemented in connection with the new fields of computer science such as evolutionary computation, machine learning, swarm intelligence, sensor technologies, and cognitive science as cognitive technology that makes the IMS-BP adaptive and self-organizing in order to face with unpredictable changes and disturbances of the manufacturing environment. The remainder of the paper is organized as follows. The next section shows an overview and the weaknesses of the existing manufacturing systems which are more or less using the bio-inspired technologies. Section 3 presents a concept of Intelligent Manufacturing System with Biological Principles (IMS-BP). Section 4 introduces bio-inspired technologies for carrying out the IMS-BP. The core aspects for achieving the IMS-BP are described in Section 5. The implementing the test-bed for adapting to disturbances within the machining system is presented in Section 6. Finally, some conclusions are given in Section 7.

2. Related work The conventional manufacturing systems, such as the

40

International Journal of CAD/CAM Vol.10, No. 1, pp. 39~50

flexible manufacturing system, computer integrated manufacturing, and reconfigurable manufacturing systems are unable to adapt to the complexity and dynamic of the manufacturing environment. These systems activate the automatic operations by using the pre-instructed programs that reduce the flexibility of the systems. The new trend of the manufacturing system development is to apply autonomous behaviours inspired from biology for the manufacturing systems. The evolution of manufacturing systems toward the intelligent manufacturing systems is shown in Fig. 1. Bio-inspired technologies have been applied for carrying out intelligent manufacturing systems in recent years. Existing researches can be classified into two groups: the evolutionary algorithms based system and the manufacturing control system. In the first group, evolutionary algorithms inspired from biology such as genetic algorithms, ant colony optimization and particle swarm intelligence are applied for the applications of CAPP [Sumichrast 2000, Wang 2005, Wang 2009, Shan 2009]. In the second group, many novel paradigms that are known as intelligent manufacturing systems were proposed in the literature. The Genetic, Biological, and Holonic manufacturing systems are the most remarkable concepts. The biological organisms have two types of information: Genetic information (DNA-type) and Knowledge information (Brain/Neuraltype or BN-type). The DNA-type information evolves through progressive generations while the BN-type information is achieved during the lifetime of one organism by learning [Brussel 1995]. Unification of information makes organisms to show functions such as self-recognition, self-growth, self-recovery and evolution [Ueda 2000]. In the Genetic Manufacturing System (GMS) [Brussel 1995], the information classification of manufacturing systems based on the information types of biological organisms was proposed. The DNA-type information includes the data of products and manufacturing equipments.

The data of products determine which processes are required by the product while the data of manufacturing equipments show which machines the corresponding processes are operated. On the other hand, the BN-type information consists of the rules for cooperating machines in order to carry out processes. The technologies for embedding this biological information to the machines and work-pieces such as Radio Frequency Identification (RFID), magnetic magnesium were proposed [Denkena 2010]. The GMS concept was improved by the recent researches in which the intelligent behaviours of the system adapted to biological organisms were added. In the Biological Manufacturing System (BMS) [Ueda 2000, Ueda 2007], machine tools, transporters, robots, and so on should be seen as biological organisms, which are capable of adapting themselves to environmental changes. In order to realize BMS, agent technology was proposed for carrying out the intelligent behaviours of the system such as self-organization, evolution and learning [Ueda 2006]. The reinforcement learning method was applied for generating the appropriate rules that determine the intelligent behaviours of machines. However, this system only achieves a greater efficiency if the agents are equipped with the cognitive capabilities, which improve the autonomous behaviours of the agents. Currently, the cognitive technology which equips the cognitive behaviours for the system such as perception, reasoning, and action is applied independently with the agent technology in the manufacturing applications [Zhao 2008, Zaeh 2009]. In the Holonic Manufacturing System (HMS), the ADACOR holonic manufacturing control architecture was proposed [Leitao 2008]. In this architecture, the manufacturing control architecture is divided into holons [Tharumarajah 1998, Christo 2007] such as the product, task, operational, and supervisor holon [Leitao 2008]. Operational holons represent the physical resources

Fig. 1. Evolution of manufacturing systems toward the intelligent manufacturing.

Hong-Seok Park and Ngoc-Hien Tran

available on the shop floor. These holons adapt to unexpected disturbances such as the machine breakdown, tool wear and so on by themselves or by the interaction with other operational holons through a supervisor holon. The mechanism for adapting to disturbances is based on pheromone techniques inspired from the ant colony optimization. The pheromone parameter indicates the level of impact of the disturbance. In this architecture, there still exists the weakness of traditional centralized and sequential manufacturing systems due to the use of the supervisor holon that reduces the flexibility of the system to respond to disturbances. This weakness will be overcome by a decentralized control architecture in which the agent technology is applied for implementing the logical part of operational holons so that these holons can directly interact among them for overcoming disturbances [Matsuda 2006]. The integration of agent and cognitive technologies for implementing the intelligent manufacturing systems is necessary. Each agent equipped with cognitive capabilities such as perception, reasoning and action, which is called cognitive agent, is smarter than a traditional agent [Park 2010]. The scope of autonomous activities of traditional agents is reduced by using the rules, while the cognitive agents use the reasoning mechanism for decision making. Most current researchers focused on the agent technology [Monostori 2006], and the evolutionary algorithms [Wang 2005] in the separate manufacturing applications, while only a few researchers concentrated on the cognitive technology [Zhao 2008]. The integration of these technologies brings a greater efficiency for manufacturing applications [Xiang 2008, Park 2010]. The contribution of this research is the proposal of new IMS-BP concept in which the swarm of cognitive agents is used for controlling resources on the shop floor. These agents are built based on the integration of agent and cognitive technologies. Cognitive agents use

Fig. 2. Model of IMS-BP.

An Intelligent Manufacturing System with Biological Principles 41

evolutionary algorithms for generating their own process planning, and cooperate for generating whole schedule of the system. The advantages of the existing concepts are inherited and integrated into the IMS-BP concept. In which the agent characteristic, cognitive capabilities, information classification and store, and learning mechanism are adapted to the previous concepts such as Genetic, Biological, and Holonic Manufacturing Systems.

3. Concept of IMS-BP The model of IMS-BP is shown in Fig. 2 in which each resource on the shop floor is an autonomous entity which integrates information into its physical part. The autonomous characteristic of entities is shown by self-X functions such as self-organization, self-decision making, self-maintenance, and so on. This characteristic is implemented by applying a synthesis of cognitive and agent technologies to the execution level of IMS-BP. Each entity on the shop floor is controlled by a corresponding cognitive agent [Leitao 2002, Monostori 2006]. The work-piece stores DNA-information, which is an optimal route through resources inherited from the previous generation. The work-piece knows which resources are used with the determined sequence. The first generation of the optimal process planning is generated by the cooperation of cognitive agents. Each machine agent generates its own process plan by using evolutionary algorithms based on biological principles such as genetic algorithms, ant colony optimization algorithm. The machine agents negotiate the whole process planning of the system for avoiding the conflict jobs among them. The resource agents, which are the machine agents, robot and transporter agents, communicate with the work-piece agent to carry out their required operations. The DNA-information of the resource agents is the information about their capabilities while the BN-

42


information is that about their operations for the machined product received from the optimal process planning. In the case of no unexpected disturbances or changes on the shop floor, a raw material grows up to the final product by the operations of combining the DNA and BN-information which are generated from the first generation. Otherwise, the cooperation of cognitive agents is carried out for finding out an appropriate route in order to keep the manufacturing running. The data of this route and the processing parameters are used to get the final product and stored for the next generation. This idea is implemented by applying the Bio-inspired technologies, which are the Swarm Intelligence, Agent technology, Machine Learning, and Cognitive technology. Besides, the Information and Communication Technology (ICT) infrastructure such as wireless communication, RFID and Ubiquitous Sensor Network (USN) plays an important role for implementing the IMS-BP [Serrano 2007, Kim 2007, Gunther 2008].

4. Bio-Inspired Technologies 4.1 Swarm Intelligence In the natural environment, a collective intelligence is carried out by simple interactions of individuals. A concept found in the colonies of insects, namely swarm intelligence, exhibits this collective intelligence [Leitao 2009]. Swarm intelligence is established from simple entities, which interact locally with each other and with their environment. In ant colonies, the collective intelligence is given by interactions of individual ants with the limited cognitive abilities through chemical substances called pheromones as shown in Fig. 3. One of the collective intelligence of an ant colony is finding the shortest route from the nest to the food source. The first ant finds the food source and returns to the nest leaving behind a pheromone trail. Over time, the evaporation of the pheromone trail begins that reduces its attractive strength. The pheromone evaporation is a criterion for avoiding the convergence to a locally optimal solution. Ants can follow many possible ways from the nest to the food source and back again, but the strengthening of the route making them more attractive is the shortest route [Leitao 2009, Garg 2009]. Transferring this principle to the manufacturing field, most current manufacturing applications use the Swarm Intelligence technology for the separate applications such as manufacturing scheduling, manufacturing control. In which, swarm intelligence technology is expressed either as evolutionary algorithms (Ant colony optimization,

Fig. 3. The shortest route chosen by ants.

Particle swarm intelligence) [Anghinolfi 2007] or as a multi-agent system [Leitao 2002]. In order to adapt with the dynamic evolution of environment, a swarm of ants needs the self-organization ability. Self-organization is carried out by re-organizing its structure through a modification of the relationships among entities without external intervention. In manufacturing systems, which is seen as a community of autonomous and cooperative entities, self-organization is carried out by locally matching between machine capabilities and product requirements [Nakano 2007, Leitao 2008, Leitao 2009]. Each machine has a pheromone value for a specific operation and the machine with the shortest processing time for a specific operation has the highest pheromone. In the IMS-BP concept, the swarm intelligence technology is applied for an integration of manufacturing scheduling and control in which the manufacturing control architecture is a swarm of agents. Each agent represents a manufacturing resource such as a robot, a machine tool or a work-piece. These agents use the ant colony algorithm for generating their operation planning, and then negotiate to generate the whole scheduling for the system. The embedded intelligence and learning skills for each agent determines the flexibility degree of its behaviours. In order to increase the intelligent behaviours of agents, cognitive capabilities are equipped for agents by using the cognitive technology.

4.2 Cognitive Agent From the swarm intelligence aspect, manufacturing systems are considered as a swarm that shows the collective intelligence by interactions among the holons. In order to implement the holons, agent technology is used [Leitao 2002]. A synthesis approach of the agent and cognitive technologies is applied to improve the autonomous characteristics of conventional agents. As the result, a smarter agent, namely cognitive agent, is proposed. The cognitive agent is a computer program equipped with artificial cognitive capabilities in order to perform the cognitive activities which emulate the cognitive behaviours of human such as perception, reasoning and decision making, communication, and learning [Park 2010]. In the IMS-BP, cognitive agents ensure the flexibility of the manufacturing system for adapting to the changes and unexpected disturbances. The cognitive agents implement a cognitive perception-action loop that does not only adapt to the changes of manufacturing system autonomously but also inherits any new optimal plan, generated for adapting to disturbances. Fig. 4 shows the architecture of a cognitive agent for processing events from the resource controlled by the cognitive agent such as tool wear, machine breakdown and so on. The perception module is responsible for a data acquisition from the resource through sensors. This data observation includes many kinds such as visual data, auditory data, vibration data, and so on. The interpretation module is responsible for transforming these data to the



5. Core Aspects to Achieve the IMS-BP

Fig. 4. Architecture of cognitive agent.

standard format that describes the status of the resource. The cognitive agent uses its own knowledge and experience to make a decision that is suitable for the status of the resource. In order to face with an unfamiliar status, self-learning and inheritance capabilities are equipped for the agent. Artificial Intelligence (AI) techniques are applied to generating a new solution for adapting to the unexpected status of the resource. Machine learning techniques equip the cognitive agent with learning capabilities in order to adapt with the dynamic evolution of the environment. The cognitive agent makes a decision that is a new plan based on the updated knowledge to face with the resource changes. The action module processes this plan into tasks and executes the tasks to the resource. At the same time, this module sends the tasks to the communication module if the plan is processed by the agent team.

Fig. 5. Swarm of cognitive agents for the machining system.

5.1 Intelligent components A component is called intelligence if it shows advanced characteristics, which are unique identification, communication with the environment, ability to store data about itself, language to display its features or production requirements [Meyer 2009]. Currently, the intelligent components use bar codes or RFID tags for storing data. The tags are attached to the machine components, which may be detached occasionally during the manufacturing process. Inspired from the biological organism of which the information exists within itself, the entity and information in manufacturing systems are separated. A new approach is to store directly the production data on the component surface by merging the information and component [Schmidt 2008, Denkena 2010]. The chosen approach is to develop the magnetic magnesium (Mg). The Mg used as a sintered material is integrated into an appropriate component [Wu 2008, Wu 2010]. The vision of “feeling” machine components is achieved by attaching multi-sensor system to these components [Denkena 2008]. Intelligent components are the results of applying sensor technologies and the ICT progress that ensure the precise operations and flexibility of the manufacturing system. 5.2 Intelligent Machining Process Planning A machining process planning provides information on how to machine the designed products. This information includes machining operations and parameters needed to be used for the machining system to convert a part from a raw material to a given shape. The machining process planning is automatically generated by using the CAPP system. Most of the CAPP systems available today are centralized in architecture and off-line data processing.

44


Even though AI techniques are used in order to generate an optimal process planning, the typical CAPP systems are difficult to make adaptive decisions in advance, without knowing actual the status of machines on the machining shop floor [Zhang 2007]. This paper proposes a cognitive agent-based intelligent process planning approach to generate a machining process planning for the machining system. In this method, work-piece, machine tools, robot, transporter, and so on are controlled by the corresponding cognitive agents. The machining system is considered as a swarm of cognitive agents shown in Fig. 5 in which the whole machining process planning for the designed product is generated by cooperating among cognitive agents. Each machine agent has the ability to generate a process plan for itself and negotiates with other agents in order to perform the whole process planning for the machining system. The negotiation ensures that all jobs are allocated to the corresponding machine agents, consequently, avoiding conflicted jobs among these agents. The information model for generating an optimal process planning is shown in Fig. 6. A work-piece agent stores the information of the work-piece, including material, shape, dimensions and information of machining features. This information of the work-piece agent is delivered by a Computer Aided Design (CAD) system. Each feature has one or several sets of operation types, and each operation can be carried out by one or more sets of the machine and tool. The tool approach direction is fixed for the tool acting on a feature. Therefore, an operation is one or several sets of three typical elements: machine, tool, and tool approach direction. The order of operations is set by precedence relationships between

the operations. Each machine agent can determine which operations are to be carried out by itself and generate its own schedule based on the work-piece and product information from the work-piece agent and its machine information. The machine agents communicate and negotiate with each other to develop an optimized executable schedule for the whole machining system. The negotiation for job allocation among cognitive agents uses criteria such as a precedence relationship between the operations and pheromone value. These criteria are used in order to avoid conflicted jobs among the cognitive agents, consequently, achieving the shortest processing time. The machine with the shortest processing time for a specific operation has the highest pheromone. This machine will be chosen to perform the specific operation in case this operation can be performed on alternative machines. As a result, the whole machining schedule is generated by cooperating of the cognitive agents. Through negotiations, each machine agent decides which jobs will be done by itself. When the job allocations are determined, the machine agent cooperates with the robot agent and transporter agent to machine the target work-piece. The machining schedule of the system changes dynamically when disturbances happen such as the tool wear, machine breakdown, malfunction of robot or transporter, and so on. The cognitive agents carry out a negotiation for rescheduling through wireless communication.

5.3 Intelligent Assembly Process Planning Traditionally, an assembly process planning comprises the generated assembly plans, resource schedule and system plan, which are isolated fields and performed by engineers. In order to overcome the increasing complexity

Fig. 6. Information Model of Agent-based intelligent machining process planning.


in the assembly operations, AI techniques as evolutionary algorithms are utilized for the automatic generation of assembly planning [Sumichrast 2000, Wang 2005, Wang 2009, Shan 2009]. In this paper, the integration of planning and execution level of the assembly operations is proposed that allows the assembly system to react autonomously to disturbances and changes. A synthesis of agent and cognitive technologies is used for this approach in which an assembly system is controlled by a cooperation of cognitive agents. Fig. 7 illustrates the cooperation of a part agent and assembly machine agents for assembling a product. In this system, the part agent generates an optimal assembly process planning and dispatches the assembly operations to the corresponding assembly machine agents. The assembly machines recognize the sub-parts which are assembled by them through part identification (Part ID) stored on the sub-parts. As a result, the assembled product is carried out by a cooperation of the part agent and the assembly machine agents. In order to implement the swarm of cognitive agents for the assembly system, the information model of the cognitive agents based assembly system is described in Fig. 8. In which a new assembled product, product

Fig. 7. Swarm of cognitive agents for the assembly system.

Fig. 8. Information Model of Agent-based intelligent assembly process planning.


information and precedence constraints are determined. The part agent uses the information of the product and assembly machines to generate an optimal assembly process planning. The product information delivered from the CAD system includes the information of a main part and sub-parts. The main part is based on the assembly operation while the sub-parts are assembled with the main part. The part agent gets information of assembly machines by communicating with the assembly machine agents. Evolutionary algorithms are used by the part agent for automatic generation of an optimal assembly process planning. The assembly operations are dispatched to the corresponding assembly machine agents by the part agent. The sub-parts are assembled with the main part by cooperation of the part agent and assembly machine agents based on the determined assembly sequence.

5.4 Intelligent Maintenance System Requirements for an attributive or functional maintenance in manufacturing systems are based on the attributes or functions of each machine or an entire system. In attributive maintenance, the manufacturing system must be stopped for replacement of faulty components or equipments. In the case of the functional maintenance, the maintenance system aims at repairing functions of the manufacturing system. A new method for functional maintenance, called self-maintenance, was proposed by Labib [Labib 2006]. A self-maintenance machine is shown in Fig. 9. In this method, the machines are equipped capabilities to monitor, diagnose, forecast and repair themselves in order to increase their uptime. Sensor technologies and AI techniques are applied to fulfil the self-maintenance function. Inprocess measurement of process parameters for the maintenance is carried out by intelligent sensor technologies. Intelligent functions such as diagnosis, forecast and decision making are equipped by AI techniques. The required capabilities of a self-maintenance machine are defined as follows [Labib 2006]: . Monitoring capability: RFID technology and USN applied in the past still show potentials for on-line

Fig. 9. Self-maintenance machine.

46


condition monitoring. Besides, new sensor technologies are proposed such as magnetic magnesium for data storage on the component in order to improve monitoring capability. . Matching capability: The machine condition is determined at a normal or an abnormal state based on the comparison of the detected status with the planned status. . Diagnosing capability: In the case of an abnormal state, the causes of faults are diagnosed and identified by applying artificial neural network or fuzzy logic. . Predictability: System operations in the past are inherited to predict failures of the system in the future [Yang 2008]. The approach of feature domain prediction is shown in Fig. 10. Features are performance signatures of the machine extracted from various sensors on the machine. If the regions of features describing the normal process behaviours and model of failures are known, then an overlap between the model of failures and the normal behaviour region indicates the predicted confidence value of failure in the future.

Fig. 10. Feature domain approach for predictability.

Fig. 11. Test-bed architecture of the intelligent machining system.

. Decision making capability: Repair actions based on the result of diagnosis and functional maintenance are proposed by the system. The repair planning action is performed by using the AI techniques such as knowledge-based systems and case-based reasoning. . Repair executing capability: The maintenance is carried out by the computer control system and actuators of the machine without any human intervention.

6. Implementation of an Intelligent Machining System In order to prove the efficiency of the proposed IMSBP concept, many jobs should be done to have the experimental results, which are the implementing agents and their mechanism for cooperation, and then the generation of the process planning by cooperating agents, and self-maintenance system. In order to implement these jobs, the agent architecture and its mechanism for cooperation should be implemented as the first job. So the test-bed of the intelligent machining system was implemented to verify the mechanism of cognitive agents for adapting to disturbances within the machining system. The architecture of the test-bed is shown in Fig. 11. In order to save the investment and installation efforts, disturbance generators (turn on/off switches) are used to generate disturbances. The Programmable Logic Controllers (PLCs) considered as the controllers of the machine tools get the process information from the Manufacturing Execution System (MES) and execute the machining jobs. The processing information of the system is displayed on the screen of the monitoring system. The work-piece information is collected by the RFID system. Each of Personal Computer (PC) includes three cognitive agents such as the machine, work-piece, and transporter agent, which are responsible for managing the machining process of a work-piece using a machine tool. Cognitive agents run on a Java Agent Development Environment (JADE) platform installed on PCs. This platform is an



agent development tool that provides a set of libraries for developing an agent-based system and an environment for running the agent application [Bellifemine 2007, Balachandran 2010]. The wireless communication is used for the interaction between the machine agents. The interaction is done by the message processing in which the eXtensible Markup Language (XML) language is used to format the messages for translating the data. The Foundation for Intelligent Physical Agents-Agent Communication Language (FIPA-ACL), an agent communication language, is used for sending messages between the agents. The negotiation of agents is implemented by using the contract net protocol [Zhou 2007]. Fig. 12 illustrates the scenario to apply to the test-bed. The machining system consists of three machine centers. The process sequence of the machining shop, the initial values of the processing time and the pheromone values of the operations (expressed by the variable pv (task #i)) are supposed, which are shown in Table 1. An assumption is the task #2 that can be done at any machines. At the beginning, the MES system dispatches the jobs to the corresponding machines based on the machine agent ID. The machining system activates the route #1 in which the machine #1 (M1), machine #2 (M2), and machine #3 (M3) carry out the task #1 (T1), task #2 (T2), and task #3 (T3), respectively. In the case of the breakdown of the machine #2, the route #2 is activated that keeps the machining system running. The mechanism for negotiating of the machine agents

is shown in Fig. 13. The disturbance occurs at the machine #2 that is shown by turning “ON” of the disturbance generator. The machine agent #2 gets the disturbance signal through the PLC #2. Immediately, the negotiation of machine agents is activated. The machine agent #2 sends a message for help to the remaining machine agents. This message content consists of the machining information and addresses of the receiving machine agents. The machine agents negotiate to find out another route. This negotiation is based on the evaluation of the pheromone values of machine agents, the precedence relationship between the operations, and current status of the machines. Each machine has a pheromone value for a specific operation and the machine with the shortest processing time for a specific operation has the highest pheromone. The machine has the highest pheromone value that is chosen for carrying out the jobs of machine #2. After negotiating, the machine agent #3 accepts the machining job of the machine #2 based on its own machine information, current status, and work-piece information. The machine agent #3 requests the scheduling information from the MES system, and then cooperates with the transporter and work-piece agent to carry out the accepted job. As the result, the system activates the route #2. The green light at the machine #3 on the test-bed is “ON”. In comparison to the current methods for adapting to disturbances, the adaptation of the machining system to disturbances based on the cognitive agents is a feasible solution. Currently, the real machining system should be stopped repairing the damaged machine, which reduces

Fig. 12. Scenario for adapting to disturbance of the machining system. Table 1. Process sequence of the machining shop Machine

Process

Processing time (min)

Pheromone value

Machine center #1 (M1)

Turning (T1)

4

pv(t1)=2


Drilling (T2)

2

pv(t2)=1


Milling (T3)

5

pv(t3)=3

(M1)

(T2)

3

pv(t2)=2

(M3)

(T2)

2

pv(t2)=1

48


Fig. 13. Negotiation process within the test-bed.

the productivity of the system by the down time of the machine tools. This method also overcomes the weakness of the holonic manufacturing system (HMS) by using cooperating directly among agents instead of using the supervisor holon in the HMS. The weakness of the centralized control still exists in the HMS by using the supervisor holon that reduces the flexibility of the system to respond to disturbances and changes [Leitao 2002].

7. Conclusions In order to face with the changes and unpredictable disturbances within the manufacturing system, Intelligent Manufacturing Systems are new trends of the modern manufacturing where the progresses of ICT, Biology, Cognitive science and novel frameworks of Distributed Manufacturing Control are integrated. This paper proposes the Intelligent Manufacturing System with Biological Principles (IMS-BP) such as an emergence of collective behaviours from the interaction of autonomous entities, self-organization and evolution. The manufacturing system has advanced characteristics such as inheritance, intelligence, and adaptation when applying these biological principles. This paper focuses on the integration of planning and execution level of the conventional manufacturing systems based on the swarm of cognitive agents inspired from the biological field. Cognitive agents increase the robustness of the system by avoiding the centralized control and show potential for implementing the autonomous behaviours by flexible ability in decision making. The core aspects in order to achieve the IMS-BP are analyzed. Genetic algorithms or the ant colony optimization is used for cognitive agent to generate its own process planning. The negotiation of agents is carried out for

generating a whole process planning of the system. In maintenance, the neural network algorithms and reasoning mechanisms equip the intelligent functions for the selfmaintenance machine such as diagnosis, forecast, and learning. The cognitive agent based manufacturing control is a feasible solution for adapting to the unexpected changes and disturbances by the advanced characteristics of cognitive agents. Besides, intelligent components applied the ICT progresses and sensor technologies ensure the precise operations and flexibility of the manufacturing system. RFID technology and new technologies for embedding information directly on components play an important role for collecting data and tracking status of moving parts such as work-pieces, transporters that ensure the processes of cognitive agents in real time. However, this concept also has a number of obstacles for applying to manufacturing. The number of standards and platforms usable for developing an agent system under industrial conditions is very limited. The existing standard for developing agent system, namely Foundation for Intelligent Physical Agents (FIPA), does not address problems related to real-time control. These obstacles require programmers to develop protocols for interaction between the developed agent systems and existing control devices. The implementation for generating the process planning through the agent cooperation, and implementing the IMS-BP concept on the real manufacturing system in order to prove the efficiency of the proposed concept will be the objects for further research.

Acknowledgements This work was supported by the 2009 Research Fund of University of Ulsan.



References

[20] Park, H.S., Tran, N.H., Song J.Y. and Kim, D.H. (2010), Cognitive Agent based Control of a Machining Shop, Proceedings of the 43rd CIRP International Conference on Manufacturing Systems, 543-550. [21] Schmidt, M., Wriggers, F.S., Fisser, F. and Nyhuis, P. (2008), Gentelligent Parts: A Decentralized Information System for Enterprises, IFIP International Federation for Information Processing 254, 125-134. [22] Serrano, V. and Fischer, T. (2007), Collaborative Innovation in Ubiquitous Systems, Journal of Intelligent Manufacturing 18, 599-615. [23] Shan, H., Zhou, S. and Sun, Z. (2009), Research on Assembly sequence planning Based on Genetic Simulated Annealing and Colony Optimization Algorithm, Assembly Automation 29(3), 249-256. [24] Sumichrast, R.T., Oxenrider, K.A. and Clayton, E.R. (2000), An Evolutionary Algorithm for Sequencing Production on a Paced Assembly Line, Decision Sciences 31(1), 149-172. [25] Tharumarajah, A., Wells, A.J. and Nemes, L. (1998), Comparison of emerging manufacturing concepts, IEEE International Conference on Systems, Man and Cybernetics, 325-331. [26] Ueda, K. (2007), Emergent Synthesis Approaches to Biological Manufacturing Systems, Digital Enterprise Technology, Springer. [27] Ueda, K. Kito, T. and Fujii, N. (2006), Modeling biological manufacturing system with bounded-rational agents, Annals of the CIRP 55(1), 469-472. [28] Ueda, K., Hatono, I., Fujii, N. and Vaario, J. (2000), Reinforcement Learning Approaches to Biological Manufacturing Systems, Annals of the CIRP 49(1), 343-346. [29] Wang, J.F., Liu, J.H. and Zhong, Y.F. (2005), A novel ant colony algorithm for assembly sequence planning, Int J Adv Manuf Technol 25, 1137-1143. [30] Wang, L. and Keshavarzmanesh, S. (2009), Assembly process planning and its future in collaborative manufacturing: a review, Int J Adv Manuf Technol 41, 132-144. [31] Wu, K.H., Gastan, E., Rodman, M., Behrens, B.A., Bach, Fr.W. and Gatzen, H.H. (2010), Development and application of magnetic magnesium for data storage in gentelligent products, Journal of Magnetism and Magnetic Materials 322, 1134-1136. [32] Wu, K.H., Traisigkhachol, O. and Gatzen, H.H. (2008), A concept for magnetic data recording on genetically intelligent (“gentelligent”) components, Physica status solidi. C. Current topics in solid state physics 5(11), 3547-3550. [33] Xiang, W. and Lee, H.P. (2008), Ant colony intelligence in agent coordiantion for dynamic manufacturing scheduling, Proceedings of the 7th World Congress on Intelligent Control and Automation, 3527-3532. [34] Yang, Z.M., Djurdjanovic, D. and Ni, J. (2008), Maintenance scheduling in manufacturing systems based on predicted machine degradation, J Intell Manuf 19, 87-98. [35] Zaeh, M.F, Beetz, M., Shea, K et al. (2009), The Cognitive Factory, In: EIMaraghy, Hoda A (ed) Changeable and reconfigurable manufacturing systems, Springer, 355-371. [36] Zhang, W.J. and Xie, S.Q. (2007), Agent technology for collaborative process planning: a review, Int J Adv Manuf Technol 32, 315-325. [37] Zhao, X. and Son, Y. (2008), BDI-based Human Decisionmaking model in Automated Manufacturing Systems, Int J of Model and Simul 28(3), 347-356. [38] Zhou, B.H., Li, C.C. and Zhao, X. (2007), FIPA agent-based control system design for FMS, The International journal of Advanced Manufacturing Technology, 31, 969-997.

[1] Anghinolfi, D., Boccalatte, A., Grosso, A., Paolucci, M., Passadore, A. and Vecchiola. C. (2007), A swarm intelligence method applied to manufacturing scheduling, Proceedings of WOA, 65-70. [2] Balachandran, B.M. (2010), Developing intelligent agent applications with JADE and JESS, Knowledge-based Intelligent Information and Engineering Systems, 5179, 236-244. [3] Bellifemine, F., Caire, G. and Greenwood, D. (2007), Developing Multi-Agent Systems with JADE, John Wiley & Sons. [4] Brussel, H.V (1995), “Navigation” issues in Intelligent Autonomous Systems, IAS-4 Conference Proceedings, 42-52. [5] Christo, C. and Cardeira, C. (2007), Trends in intelligent manufacturing systems, Proceedings of the IEEE International Symposium on Industrial Electronics, 3209-3214. [6] Cus, F. and Zuperl, U. (2009), Particle swarm intelligence based optimisation of high speed en-milling, Computational Materials Science and Surface Engineering, 1(3), 148-154. [7] Denkena, B., Henning, H. and Lorenzen, L.E. (2010), Genetics and intelligence: new approaches in production engineering, Production Management 4, 65-73. [8] Denkena, B., Mohring, H.C. and Litwinski, K.M. (2008), Design of dynamic multi sensor systems, Production Engineering 2, 327-331. [9] Garg, A., Gill, P., Rathi, P., Amardeep and Garg, K.,K. (2009), An Insight into Swarm Intelligence, International Journal of Recent Trends in Engineering 2(8),42-44. [10] Günther, O., Kletti,W. and Kubach, U. (2008), RFID in Manufacturing, Springer. [11] Kim, M., Lee, Y. J. and Ryou, J. (2007), How to Share Heterogeneous Sensor Networks in Ubiquitous Environment, International Conference on Wireless Communications, Networking and Mobile Computing, 2799-2802. [12] Labib, A.W. (2006), Next Generation Maintenance Systems: Towards the Design of a Self-maintenance Machine, IEEE Int. Conf. Industrial Informatics, 213-217. [13] Leitao, P. (2008), A Bio-Inspired Solution for Manufacturing Control Systems, In: Azevedo, A.(ed.) IFIP International Federation for Information Processing, Innovation in Manufacturing Networks, 303-314. [14] Leitao, P. (2009), Holonic Rationale and Bio-Inspiration on Design of Complex Emergent and Evolvable Systems, Transactions on Large-Scale Data and Knowledge-Centrered System I, Springer. [15] Leitao, P. and Restivo, F. (2002), Agent-based Holonic Production Control, Proceedings of the 13th International Workshop on Database and Expert Systems Applications, 589-596. [16] Matsuda, M., Ishikawa, Y. and Utsumi, S. (2006), Configuration of machine tool agents for flexible manufacturing, Proceedings of the 39th CIRP International Conference on Manufacturing Systems, 351-357. [17] Meyer, G.G., Framling, K. and Holmstrom, J. (2009), Intelligent Products: A survey, Computers in Industry 60, 137-148. [18] Monostori, L., Vancza, J. and Kumara, S.R.T. (2006), Agent-based Systems for Manufacturing, CIRP Annals Manufacturing Technology 55(2), 697-720. [19] Nakano, T. and Suda, T. (2007), Applying Biological Principles to Designs of Network services, Applied Soft Computing 7, 870-878.

50


Hong-Seok Park received his B.Sc. degree from HanYang University in Korea in 1979. He received his Dipl.-Ing. degree from RWTH Aachen and Dr.-Ing. degree from University of Hannover in 1987 and 1992, respectively, in Germany. He is now a Professor in the School of Mechanical and Automotive Engineering at University of Ulsan in Korea. His research interests include Intelligent Manufacturing System, Digital Manufacturing Techniques, and CAD/CAM/CAE.

Ngoc-Hien Tran received his B.Sc. degree from University of Transport and Communications and M.Sc. degree from Hanoi University of Technology, in 2001 and 2007, respectively, in Vietnam. He is now a PhD student in the School of Mechanical and Automotive Engineering at University of Ulsan in Korea. His research interests include Intelligent Manufacturing System and Cognitive Agent based Manufacturing applications.

Hong-Seok Park

Ngoc-Hien Tran