In order to acquire the farsighted development, the business enterprises not only need to face the complicated exterior environment, but also need to face the enterprises resources management. ...Multi-location manufacturing enterprises are being forced into greater collaboration with customers, suppliers and inner enterprise resources in order to produce quality products in smaller batches, shorter lead times and with greater variety. This paper presents a creative approach for solving dynamic planning and scheduling problems in hybrid distributed manufacturing execution system (HDMES). We adopt a facility layout optimization approach based on single genetic algorithm (SGA) and parallel genetic algorithm (PGA). Different approaches used to model various aspects of manufacturing processes are reviewed and found, and an approach based on scheduling rules, multi-agent technique, multi-closed loop control and user cooperation is presented to solve the distributed planning and scheduling problem. We hierarchically decompose the large-scale planning and scheduling problem into the planning level problems, the scheduling level problems and material tracking feedback level problems. In the optimization approach, the genetic operators and selection method are used to better transmit the temporal relationships in the chromosome, and island model PGA are proposed. The optimization approach based on genetic algorithm is tested on typical production management problems in hybrid distributed manufacturing execution system. The results are compared with other proposed approaches. Compared to traditional genetic algorithm, the proposed approach yields significant improvement in solution quality. The superior results indicate the successful incorporation of a method to generate initial population into the genetic operators
On the Control of Production Networks (PN) numerous research projects and activities have been initiated. There is a general regret about lacking theoretical backgrounds for the development of ...resulting methods and instruments. The interpretation of Production Networks as generalised topological structures allows generating a theory design, putting relevant models on a common base. In order to demonstrate the power of the approach, the planning and control problem is tackled. Employing the Self Similarity principle, the solution conjecture is valid for the organisation issues as well as the process control. It can be shown that human decision making and software agent negotiations, self similarly based on the same models, may provide simple solution procedures. The approach is motivated by fractal organisation experiences, complexity theory and topology. Diverse planning and control problem aspects and/or sub-problems are covered by sets of models constituting model systems, progressively generating solutions by network units' adaptations. The paper intends to contribute to the production systems theory's extension to the field of production networks.
Due to the widespread availability of the Internet, large scale distributed projects in manufacturing are becoming popular. In order to address the dynamic requirements of multi-project production, ...this paper develops a collaborative multi-project planning and scheduling system based on the multi-agent technology. The hybrid framework of system is presented and discussed with detail. Six types of intelligent agents are proposed in this system. These agents are dynamically deployed at each location where a project is executed. Further more, a negotiation mechanism for multi-project planning and scheduling is developed to coordinate these distributed agents. Task allocation strategy and critical resource leveling strategy are also presented . Through which a predictable and nearly optimal scheduling with agility and adaptability can be realized in distributed multiple projects manufacturing environment. Finally, a prototype system was implemented upon this framework, and was validated in an aerospace company.
With the intensive development of networks and distributed applications, industry must define new concepts of e-manufacturing and develop new solutions that generates gains of productivity. This ...paper considers the manufacturing execution system (MES) facility, and presents two approaches which implement this functionality. The first approach addresses enterprises inter-connected in a virtual factory, and focuses on the monitoring service of the distributed manufacturing activities. The second approach uses a multi-agent technology to drive the production of a flexible manufacturing system, and it is shown how agents can collaborate to carry out scheduling and resource allocation within a plant.
Human-centred technology is an alternative approach to the design of manufacturing systems where the operator is considered an "asset" rather than a "nuisance". A human-centred system utilizes and ...relies on the operator's skill, knowledge, flexibility and creativity. It also gives him the control over and the responsibility for the system. This paper gives a short introduction into human-centred technology and illustrates the concept of human-centred CIM by outlining the project on a "Human-Centred Machining Cell" at Imperial College. In addition it points to related projects which have either just finished, such as the ESPRIT 1217(1199) project on "Human-Centred CIM" or in which the work started earlier is being continued. The paper does not present completely new findings or developments, but it attempts to summarize and update the present situation in a number of projects in the field without claiming to be comprehensive.
Contract nets, which use a bidding mechanism to solve problems, were first proposed as artificial intelligence constructs for distributed problem solving. They have recently been applied to ...distributed manufacturing control. The contract net presented, called a complete contract net, is intended to implement both high- and low-level control. The complete contract net incorporates mechanisms for just-in-time manufacturing, dynamic lot sizing, and bids based on known costs rather than surrogate costs. Also presented is AdaBuy, a simulator for complete contract nets, and the results obtained using it.< >