Chin-Yin Huang

Distributed manufacturing execution systems: A workflow perspective

Research of manufacturing execution systems (MESs) has been conducted to integrate ERP and physical operational systems, so a fully automated and integrated manufacturing management environment can be developed. However, as subcontracting becomes common in industries, a single MES is not enough and it is necessary to develop distributed MESs to integrate the distributed operations. This research proposes a distributed workflow model to develop such a system. In the workflow model, there are three cooperation types for defining the relationships between an action and resources. Six relational patterns are defined for the relationships between actions. This research also reviews current information technologies, including data acquisition and control (DAC) components, control softwares, ASP .NET, and CLIPS, for supporting the development of distributed workflow models. Additionally, this research proposes using a collaborative workflow-based protocol (CWP) to form a distributed workflow. CWP is based on the integration of Internet technology and CLIPS. It is found that the proposed structure may ease the decision process of forming a distributed workflow. In summary, this article reports a cutting edge approach for the development of distributed manufacturing execution systems. The results show solutions for integrating high-level planning systems, such as ERP, and low-level operational systems through information technologies.

Enterprise agility: a view from the PRISM lab

The agility of enterprises is analyzed in this article from two perspectives: business and organizational agility, and operational and logistics agility. Because of the powerful support of information technologies, mainly AI and networking, companies can seek collaborators to accomplish complex customers’ requirements without investing to expand their own capacity. The emerging alliances of enterprises are virtually formed for various customers’ and markets’ needs. This concept sustains business and organizational agility. In terms of operational and logistics agility, this article suggests that the connection between the autonomy functions and agility requires further study. Our research has proposed that error detection and recovery, and conflict resolution are two significant functions of operational and logistics agility, and determine the expected benefits from the business and organizational agility. The link between a given enterprise flexibility and agility is also analyzed.

Agility of networked enterprises — parallelism, error recovery and conflict resolution

Abstract Globalization has transformed companies to enterprises over the last 50 years. Computing and communication have also transformed production and service organizations over the past 50 years. While parallelism, error recovery, and conflict resolution have been addressed by human workers since the early days of industry, they have recently been transformed by computer-integrated functions. A major concern in the global market is agility. The agility of enterprises is considered an increasingly important factor of economic competitiveness. We view it from two perspectives: business and organizational agility, and operational and logistics agility. In the business and organizational perspective of agility, our research has developed an analytic method called distributed parallel integration evaluation model (DPIEM). Its purpose is organizing/reorganizing resources among distributed, networked organizations, based on the parallelism theory of computing and communication. In terms of operational and logistics agility in such distributed organizations, our research has suggested that the connection between the autonomy functions and agility requires significant functions of error detection and recovery (EDR), and conflict resolution (CR). The impact of both functions on the operational and logistics agility is analyzed and illustrated. The article concludes with several theoretical observations about the role of information technology (IT) in modern, distributed and networked enterprises.

Autonomy and viability-measures for agent-based manufacturing systems

An Agent-Based Manufacturing System (ABMS), represented as a model of future manufacturing systems, is introduced in this article. In an ABMS, computerized agents are autonomous and working collaboratively through the regulation of structured dialogues. The autonomy functions of an agent are demonstrated by four internal functions: internal resource management, a reflexivity mechanism, a goal adjustment mechanism, and collaboration management. The four functions assist an agent in interacting with other agents and with the environment and to adjust itself to the situation in the dynamic environment. Agents are similar to living systems. Based on the concept of living systems, autonomy and viability at the agent level are proposed and defined as new measures for manufacturing systems. A theory of coexistence and parasitism of ABMS is developed to identify the coexistent and parasitic characteristics of ABMS. Applications of the theory to real cases are also illustrated. Finally, a road map for the future research of distributed, collaborative manufacturing organizations is shown, based on the ABMS model developed.

Multi-tier and multi-site collaborative production: Illustrated by a case example of TFT-LCD manufacturing

To seek collaboration with other partner enterprises becomes an inevitable strategy in todays highly dynamic and complex business. Production planning over a collaborative network relies on resolving three challenges: (1) different objectives and autonomy of partner enterprises, (2) communication and coordination among the enterprises, and (3) planning over multiple tier enterprises and multiple local plants (sites) inside an enterprise. To resolve the challenges, this paper proposes a distributed production planning system for a multi-tier and multi-site production system by combining agent technology with advanced planning and scheduling (APS) system. Two types of agents are designed: one for each tier enterprise (tier agent) and the other for each local plant (site agent) inside a tier enterprise. While a tier agent is responsible for finding a suitable plan for the demand orders by transmitting messages through designated protocols with its downstream and upstream tier agents, a site agent is designed for detailed production schedules for its local plant by interacting with its tier agent and its APS system. This paper has developed various protocols to integrate the partner tier and site agents to construct a distributed collaborative production system. This paper takes TFT-LCD manufacturing as an illuminative example to demonstrate the feasibility of the proposed multi-tier and multi-site collaborative production. The results show the feasibility of synthesizing multiple production plans over multiple enterprises and multiple production sites. Within the planning system, each company maintains their autonomy of production decision. Besides, the supply network is reconfigurable because the participated companies only need to comply with the protocols without changing their local legacy production planning systems.

Decision model for partnership development in virtual enterprises

A virtual enterprise has been considered as a temporary consortium of companies. Recently, researchers have investigated virtual enterprises from various perspectives, such as business activities and workflows, information exchanges and their standards, organizational behaviours between companies, etc. However, real collaborative activities among companies in a virtual enterprise cannot be done before a partnership is developed among companies. Besides, though partnership is important in virtual enterprises, a decision model for partnership development is seldom studied. This research develops a decision model from a micro view (i.e. from a companys perspective). A company may use the model to specify its current partnership strategy and determine its future partnership strategy with another company. Four factors are considered in the model: the format of the relationship with suppliers, the number of suppliers, the type of service provided by suppliers and the method of service delivery from suppliers. However, a partnership strategy has to be updated based on internal and external changes. This research finds out that six factors are associated with a change on partnership. With those factors, a decision process is developed for a company to update its partnership strategy. The partnership of IBM to Acer is also investigated to illustrate the application of the model. The results validate the applicability of the decision model.

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 diffficult 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.

Order confirmation mechanism for collaborative production networks

To avert risk, todays companies perform business with little capital investment. They invest only on the core activities which gain higher profit and outsource non-critical production activities to their partner companies (subcontractors or sourcing companies). Hence, a customer order is normally completed by a collaborative production network formed by the partner companies. However, co-ordinating those loosely and distributed companies to meet the uncertain demands from the final customers is not simple. This research suggests an order confirmation mechanism that allows synchronizing the available capacities over the supply network. The mechanism is developed based on a factory agent system and a design of three protocols: (1) protocol for normal condition, (2) protocol for abnormal condition, and (3) protocol for cancellation. Based on the protocols, messages and procedures are designed. The design is also deployed on a distributed computing environment to illustrate the feasibility and applicability.

Development of integrated models for material flow Design and control – a tool perspective

Abstract Modern computer and communication technologies can improve the quality, accuracy, and timeliness of decision-making, and can also provide better design tools for material flows in manufacturing systems. In this article, a tool perspective is applied to study models of material flows. Two integrated tools are introduced: facility description language (FDL) and concurrent flexible specification (CFS). The two tools represent the newer perspectives which are not considered in traditional facility modeling approaches. FDL is implemented in a 3-D emulation environment, which provides an integrated modeling tool for distributed engineers working on various aspects of manufacturing system design. In the models of CFS, the designers can develop the physical layout, functional model, information model, and control signals in an integrated modeling environment. Based on our study, it is anticipated that integrated modeling approaches with a tool perspective will continue to evolve. Interesting challenges include the development of learning mechanisms and computer-supported collaborative modeling tools.

Communication Protocols for Collaborative Forecasting

In this research, protocols for activities of collaborative forecasting among the partnered companies are developed in accordance of twelve scenarios derived by market types (excess supply or excess demand), types of focal company (retailer, manufacturer, or material supplier), and type of integration (integrated or disintegrated). Each of the twelve scenarios has its specific protocol in modified sequence diagram. Additionally, eleven protocol modules are developed for constructing the twelve protocols. With the protocol modules, the protocols are presented in a more structural format and are more readable. A prototype system in Java program is also developed to illustrate the applicability of the protocols. By deploying the system over internet, the twelve protocols are realized and can regulate the communications of collaborative forecasting.