Agent-based Manufacturing Systems Research Articles

Operating minimally intelligent agent-based manufacturing systems across the Average demand Interval – coefficient of variation (ADI-CV) demand state space

ABSTRACT Minimally Intelligent Agent-Based Manufacturing concerns the provision of agents with the minimal intelligence required to autonomously negotiate and broker work across machines and jobs. The minimal intelligence of agents can be updated in real-time and when coupled with technologies, such as Additive Manufacturing (AM), robotics, automated inventory, and Computer Numerical Control (CNC) affords highly flexible, re-configurable, resilient and responsive systems. The concept has become topical as changes to global supply chains are necessitating a shift to responsive and resilient on-demand manufacturing. However, understanding and characterising how these systems operate under various demand profiles is required to support operators operating these future systems. This paper reports a numerical study into the operating behaviour of minimally intelligent agent-based manufacturing systems operating across the Average Demand Interval and Coefficient of Variation (ADI-CV) demand state space. An established state space used in spare part supply chain research. The results show minimally intelligent manufacturing systems are stable across much of the ADI-CV demand space 85%, 63% and 76–84% for constant, triangular and lognormal manufacturing time distribution profiles respectively. The stability is largely independent of the combination of intelligence. Rather, the combination of intelligence impacts job Time-in-System and system response.

Required parameters for modelling heterogeneous geographically dispersed manufacturing systems

COVID-19 and global crises/events are driving governments to rethink their national manufacturing strategies. The drastic change of societal conditions has exposed our reliance on a constrained set of production practices. Furthermore, the future manufacturing landscape indicates - supply chain crises, trade agreements and natural disasters - a high level of volatility which requires a response that is far from being achieved.While these emergent challenges have called the efficacy of established practices into question, new manufacturing technologies, such as Additive Manufacturing (AM), present the capability to provide a solution. One proposal is agent-based brokering of AM which could be a method for tackling local, regional, national, and international production needs. However, to achieve the reality of brokered AM, it is imperative that the diversity of AM capability is considered. Diversity that existing homogeneous modelling of AM and manufacturing systems rarely consider or capture. This paper conceptualizes the reality of AM systems and elucidates parameters that are necessary for successful modelling and subsequent co-ordination. Having presented the required parameters the paper continues to discuss requisite levels of abstraction, suitable performance metrics and the role of humans in agent-based manufacturing systems.

Comparison of Three Agent-Based Architectures for Distributed Additive Manufacturing

Models of distributed agent-based Additive Manufacturing (AM) systems have suggested that considerable productivity, reactivity, and resiliency gains can be realised. However, there remains scepticism and real-world validation is required to demonstrate the benefits, practicalities, and challenges of implementing agent-based manufacturing systems. It is only then that potential gains can be evidenced and the Return on Investment fully appraised.This paper discusses the underlying communication architecture required for the creation of a hardware technology demonstrator that enables the evaluation of distributed, agent-based, AM. Three architectures to create the desired network (Host-Client, Peer-to-Peer, and Digital Shadow) are discussed, with a focus on implementation and operational factors. The paper continues by describing the future development needed for the creation of a demonstrator aimed at exploring the ease of adoption of agent-based approaches and the potential for integration into existing AM workflows. The planned demonstrator is to be implemented at the University of Bristol to co-ordinate production across multiple sites in order to support their education and research activities as well as provide a platform for public and industry engagement.

Predicate-Based Model of Problem-Solving for Robotic Actions Planning

The aim of the article is to describe a predicate-based logical model for the problem-solving of robots. The proposed article deals with analyses of trends of problem-solving robotic applications for manufacturing, especially for transportations and manipulations. Intelligent agent-based manufacturing systems with robotic agents are observed. The intelligent cores of them are considered from point of view of ability to propose the plans of problem-solving in the form of strategies. The logical model of adaptive strategies planning for the intelligent robotic system is composed in the form of predicates with a presentation of data processing on a base of set theory. The dynamic structures of workspaces, and a possible change of goals are considered as reasons for functional strategies adaptation.

Design of scalable agent-based reconfigurable manufacturing systems with Petri nets

ABSTRACTReconfigurable manufacturing systems (RMS) provide a flexible paradigm to deal with frequently changing demand and technologies. Scalability is an important property that determines whether the capacity of RMS can accommodate fluctuation in product demand by adjusting/reconfiguring machines and/or structure of production processes. Although there are several studies on scalability of RMS, there are relatively few studies that provide a framework that supports the development of scalable RMS based on multi-agent systems (MAS), from modelling, analysis to design. Motivated by this urgent need, this paper attempts to bridge the gap between theoretical development and design of scalable agent-based RMS. Although MAS and agents’ characteristics of autonomy provide a suitable architecture to model and capture interactions of entities in agent-based manufacturing systems, the original MAS lack process models for agents. By adopting Petri nets as the process models of agents, the capacity scalability problem (CSP) in agent-based RMS can be described by Petri nets in MAS architecture. Due to the lack of optimisation theory developed for Petri nets, the CSP is transformed into an optimisation problem that can be solved by applying classical Lagrange relaxation optimisation theory to relevant agents in the problem-solving processes. Therefore, the methodology proposed to develop scalable agent-based RMS is based on MAS, Petri net models and Lagrange relaxation optimisation theory. The effectiveness of agent-based RMS is illustrated by examples which show that fewer configurable resources are required for agent-based RMS to meet order requirements.

Integrating Human Operators into Agent-based Manufacturing Systems: A Table-top Demonstration

To stay competitive in the modern market, manufacturing systems must be adaptable and flexible to deal with challenges such as machine breakdowns or requests for new, customized products. A strategy proposed to improve system flexibility is the integration of multi-agent control on the plant floor. The multi-agent control strategy consists of a system of distributed agents (e.g. resource agents, product agents) that are responsible for controlling various parts of the manufacturing system through cooperation. While previous Multi-Agent Systems (MAS) implementations have proposed architectures for resource and product agents, they do not specify how human workers can be integrated to improve system flexibility. This work specifies a Human Resource Agent (HRA) architecture that can be integrated into a manufacturing system with a multi-agent control strategy. The proposed architecture is demonstrated using a table-top manufacturing testbed showcasing the flexibility the HRA can add to a manufacturing system.

The concept of intellectual manufacturing agents and the specifications of its implementation

The architecture of agent-based manufacturing systems is considered, a conceptual model of a manufacturing agent (MA) is described. The required operation features of manufacturing agent systems is self-learning, self-organization and self-adapting. The purpose of multi-agent systems is achieved by the interaction of manufacturing agents. In practice, it is offered to implement a manufacturing agent as an intelligent robotic system, which can move within the manufacturing area, carry out transport, auxiliary and separate manufacturing operations.One of the peculiar features of the MA operation at the workspace is the determination of the process equipment position and indetermination of the transport system position, other agents and humans. Therefore, a peculiar feature of the manufacturing agent control system should be the adaptability of a decision-making system, which can create and, if necessary, modify plans of the MA operation in conditions of changing the workspace of a flexible integrated manufacturing. It is offered to observe the MA workspace by using advanced sensor systems. Herewith, one of the methods of setting assignments for the MA can be the setting by visual methods and thus, the implementation of the MA visual control. It should take into account the possible multi-zone nature of MA workspaces. The paper results were used in developing the intelligent robotic system of assembling

A closed-loop feedback simulation for RFID-based manufacturing planning and control system

Nowadays, the performance of customer service level in manufacturing industries highly depends on an effective manufacturing planning and control system MPCS. An agent-based manufacturing planning and control system AMPCS allows negotiation-based decision-making to achieve the global objective. With the advantage of radio frequency identification RFID technology, agents in AMPCS may effectively control and identify the production progress of all raw materials, work-in-progress WIP, and products across the entire manufacturing supply chain. However, the production schedule may become inefficient or even infeasible due to uncertainties, such as machine breakdowns, processing time delays, rush order, quality problems, and unavailable material. Therefore, this research proposes a closed-loop feedback simulation CLFS approach based on RFID-enabled real-time manufacturing information for scheduling and rescheduling in an agent-based manufacturing system in response to the dynamically changing manufacturing activities and exceptions. From the simulation experiment, the scheduling results proposed by RFID-based MPCS combined with CLFS performed much better than the results obtained from previous researches.

Agent-Based Decision Support and Simulation for Wood Products Manufacturing

A rough mill is a manufacturing plant where lumber of approximate dimensions is cut into components of specific sizes, priorities, and qualities to fill customer orders for wood products such as furniture, doors, and window frames. Lumber is a valuable natural resource that is a significant expense to the company. By improving the processes in the rough mill, cost can be reduced and waste of natural materials is decreased. We present an overview of research in agent-based manufacturing systems. The operations in a rough mill are described and the decisions that operators take are identified. A rough mill decision support and simulation system is designed and implemented. An agent ontology for rough mill operations is developed. A prototype system is implemented to demonstrate the architecture and interagent communication. This prototype is extended to two lines of production and the negotiation protocol is presented. Extension of the approach to multiple lines of production is discussed. The prototype system is used to implement a decision support and simulation system that is validated with historical data. Finally, we present a comparison of the advantages and disadvantages of using the multiagent paradigm in rough mill decision support.

Disturbance management design for a holonic multiagent manufacturing system by using hybrid approach

This paper proposes a disturbance management methodology for an agent-based holonic manufacturing system by using hybrid control approach, with its application to a mobile manipulator system. Firstly, the architectures based on holarchy and agents are given. Then the framework of hybrid control strategy is outlined and the stability analysis for the closed-loop system is introduced. The major contribution of this work is the exploration for using hybrid control approach into disturbance management mechanism. The switched disturbance detector is presented and the identification algorithm is given. The hybrid controller is designed to reject the disturbance by switching between reactions against diagnosed symptoms. Finally, the case study onto the mobile manipulator hybrid system verifies the effectiveness and applicability of this design method. It validates the agility, efficiency, and retaining stability to system of the holonic multiagent concept for industrial or other application systems.

Manufacturing and Control: Putting Agents to Work

A review of Agent-Based Manufacturing and Control Systems by Massimo Paolucci and Roberto Sacile

Evaluating the Prometheus methodology through a case study on designing an agent-based holonic control system

An agent-based manufacturing system is a new breed of industrial control environment where autonomous software entities, called agents, cooperate to manage the hardware in a factory in a decentralised and intelligent manner. These control systems allow new product families to be introduced on the fly and empower the machinery to late-customise products. This paper analyses the applicability of Prometheus, one of the popular methodologies for designing general-purpose agent-based systems. The paper examines how useful Prometheus is for designing agent-based industrial control systems. An industrial-strength, flexible material-handling environment provides the experimental test-bed for this analysis. The key conclusion is that this methodology is very suitable for developing static interactions between agents, but lacks sufficient power to handle several of the requirements associated with a real-world industrial environment, e.g., openness of new agent types and services, system development using bottom-up design approaches and dynamic reconfiguration of the manufacturing processes.

Applications of agent-based systems in intelligent manufacturing: An updated review

Agent technology has been recognized as a promising paradigm for next generation manufacturing systems. Researchers have attempted to apply agent technology to manufacturing enterprise integration, enterprise collaboration (including supply chain management and virtual enterprises), manufacturing process planning and scheduling, shop floor control, and to holonic manufacturing as an implementation methodology. This paper provides an update review on the recent achievements in these areas, and discusses some key issues in implementing agent-based manufacturing systems such as agent encapsulation, agent organization, agent coordination and negotiation, system dynamics, learning, optimization, security and privacy, tools and standards.

Engineering framework for agent-based manufacturing control

MAS (multi-agent systems) and HMS (holonic manufacturing systems) are enabling the vision of the Plug & Play Factory and paving the way for future autonomous production systems. This paper reviews the state of the art in implementations of agent-based manufacturing systems, and identifies the lack of engineering tools as a technological gap for widespread industrial adoption of the paradigm. The lack of tools limits the implementation of agent-based manufacturing systems within reach of only a handful of domain experts. One of the current challenges for the design and implementation of intelligent agents is the simulation and visualization of the agent societies. This issue is significant as soon as the software agent is embedded into a mechatronic device or machine resulting in a physical intelligent agent with 3D-mechanical restrictions. These mechanical restrictions must be considered in the negotiations between agents in order to coordinate the execution of physical operations. This paper presents an engineering framework that contributes towards overcoming the identified technology gap. The framework consists of a comprehensive set of software tools that facilitate the creation, simulation and visualization of agent societies. The 3D framework is innovative in fully emulating the deployed agents, recreating multi-agent negotiations and societies that coordinate and execute control of assembly operations. The documented research describes the methodology for the 3D representation of individual physical agents, the related identified objects present in the interaction protocols, and the assembly features and clustering algorithms.

Comparison of negotiation protocols in dynamic agent-based manufacturing systems

This paper proposes a negotiation methodology based on multi-agent system for heterarchical and complex manufacturing control systems. This approach has been selected to implement new paradigms based on “co-opetition=co-operation+competition” in order to improve the “production on demand” and reaction capabilities of distributed production systems related to the net-economy. Agents may represent products and resources of the system. The local scheduling and control functions in dynamic environments is addressed by a new negotiation protocol between agents based on the “request session” principle for cooperation and on the game theory approach for competition.

Web-based distributed manufacturing control systems

With the evolution of the application of web technology, agent-based manufacturing systems can be both easily implemented and become a potential approach for modern manufacturing enterprises. In this paper, we present problems and models required for the development of web-based distributed manufacturing control systems. The specific objectives are: (1) to propose an appropriate architecture for a web-based distributed manufacturing control system, (2) to define the specification of a web-based cell controller, (3) to design the co-operative mechanisms, contract net protocol, for better system performance, (4) to upgrade the communication capability of existing CNC machines from RS232C to TCP/IP (Transmission Control Protocol/Internet Protocol) based on Ethernet .

Solving the n-job 3-stage flexible flowshop scheduling problem using an agent-based approach

In this paper, a general methodology of agent-based manufacturing systems scheduling, incorporating game theoretic analysis of agent cooperation is presented to solve the n-job 3-stage flexible flowshop scheduling problem. The flowshops are flexible in the sense that a job can be processed by any of the identical machines at each stage. Our objective is to schedule a set of n jobs so as to minimize the makespan. We perform error bound analysis using the lower bound estimates developed in the literature as a datum for comparing the agent-based scheduling solutions with other heuristic solutions. The results of the evaluation show that the agent-based scheduling approach outperforms existing heuristics for the majority of the testing problems.

Coordination and control in distributed and agent-based manufacturing systems

This paper addresses two fundamental issues inherent to every manufacturing system. These issues concern the coordination and control of manufacturing activities. In doing so, many aspects are reviewed and analysed. First, the main distributed manufacturing paradigms are briefly introduced. Next, the various forms of interdependence that commonly constrain manufacturing activities are reviewed, as well as the coordination mechanisms used to manage them. Then, after analysing different approaches developed to address coordination and control of manufacturing activities in distributed and agent-based manufacturing systems, this paper highlights the limits of the existing classification schemes for coordination mechanisms and provides a new scheme that extends beyond those previously reviewed in order to take into account recent advances. The studied distributed manufacturing paradigms are finally compared in terms of coordination and control.

Agent-based modeling and mapping of a manufacturing system

Considering the agent-based modeling and mapping of a manufacturing system, some system models are described in this paper, which include: domain-based hierarchical structure (DBHS), cascading agent structure (CAS), proximity relation structure (PRS), and bus-based network structure (BNS). In DBHS, one sort of agents, called static agents, individually acts as domain agents, resources agents, user interface agents and gateway agents. And the others, named mobile agents, are the brokers of task and process flow. Static agents representing a sub-system may itself be an agent-based network and should learn, as the mobile agents, to deal with new situations. Mobile agents move around the network domains taking advantage of the resources to fulfill their goals. In CAS, the authors used unified modeling language (UML) to build up an agent-based manufacturing system. It is said that the enterprise agent (the main agent) has factory agents together with some directly jurisdictional workshop agents, cell agents, and individual resource agents. Likewise, the factory agent has workshop agents together with some directly jurisdictional cell agents and individual resource agents, and so on. In PRS, the resources agents are located together by function and abilities. There is only one agent that behaves as the task-announcer. The communication just occurs among the proximity relational agents. BNS is very similar with human society, being connected with a network, some agents, such as “cost calculating”, being to just cope with matter-of-fact jobs. Some agents run as individual resources that can negotiate with each other and advertise a necessary message within the whole domain or a given group of agents. The administration just relies on the individual address of agents and the group ID code of agents.

Evaluation of agent-based manufacturing systems based on a parallel simulator

Recent research in artificial autonomous agents has paved a way to fulfill the requirements of autonomous operations in manufacturing systems in terms of planning, scheduling, control, and processing. Those manufacturing systems are also coined agent-based manufacturing systems (ABMS). However, it still lacks an evaluation tool to judge the performance of a designed ABMS. Hence, an ABMS is difficult to be accepted by users and managers, and is difficult to be realized in the real world. Based on a developed ABMS model, this research develops a simulator on a parallel computer system. The simulator is designed to imitate the changes of the environment by the changes of incoming tasks, and to imitate the internal changes by the resource breakdowns. Besides, the autonomous, communicative, and cooperative behaviors of agents are also developed in the simulators. Cases and scenarios for ABMS are developed to apply in the simulator. The analytical results show that through the simulators, decision-makers can clearly observe the viability of each agent as well as the whole ABMS. Thus, adjustments on the ABMS design can be made accordingly and the final ABMS can successfully survive in the real world.