Research On Process Preparation Resource Management In The Grid

With the trend of complication, decentralization, labilizationand intellectualization of modern enterprise enviroment, advanced manufacturinhg pattern, such as Networked Manufacturing (NM), Virtual Manufacturing (VM), Agile Manufacturing (AM), Global Manufacturing (GM) and so on, were proposed successively. These patterns have functioned varyingly in manufacturing field, which contribute greatly to manufacturing industry informatization. Due to the coincidence between the characters and supports of grid and the requirement of networked manufacturing, it is of great theoretical value and realistic significance to contruct grid manufacturing system to support global resource share and collaboration with grid technology intrduced into manufacturing field. This paper focuses mainly on the architecture and resource management of grid-based process preparation. The main research contents in this paper are as follows:Firstly, the earlier work of the research subject, digitization agile process preparation system (DAPPS), is introduced as well as its overall framework, function partition and system feature. After analyzing the problems of DAPPS's operation, requirement analysis on the application of grid in DAPPS are made. Then the conception of process preparation system grid (PPSG) is put forward and six-layer architecture of PPSG is built up. The application process of PPSG is derived and key technologies and the importance of resource management of PPSG implementation are discussed.Secondly, the search and location of grid process preparation resources is studied. The classification and description of process preparation resources in the grid are researched and a hierarchical grid resource search algorithm is proposed. After analyzing the small world feature of PPSG, VOs are constructed according to the resources' type and resources in VO are then clustered dynamically based on the resource attributes. In the clustering algorithm proposed, an optimization model of resources clustering evaluated by internal clustering degree and external detached degree is proposed based on fuzzy clustering tree with multi-λand genetic algorithm is applied to optimize the clustering results and corresponding genetic operations are studied. Furthermore, a novel approach matrix crossover is designed to prevent premature. Tthe simulation results show the effectiveness of the proposed clustering algorithm. After resource being clustered, a seach policy of resource cluster center priority and grade diffuse is proposed to seacr resources from a global view. Thirdly, multi-attribute fuzzy optimal selection of grid resources based on information axiom and rough set is studied. Combination with grid environment and features of grid process preparation resources, five metric of QoS and their quantification are given. A multi-attribute fuzzy optimal selection model of grid resources based on information axiom and rough set is proposed. This model combines the theory of fuzzy mathematics with the information theory of Shannon. Grid resources are evaluated in quantity and the optimal resource is chosen. At the same time, the weight of each attribute is determined by the dependency degree of rough set applied into optimal selection.Fourthly, QoS based resource economic scheduling in PPSG is studied.After analyzing the research on grid resources management architecture and the advantages of economic model, the market economic model is introduced into PPSG resources allocation, which divides the resource allocation into two sub questions: determination of resouce's balanced price and resource scheduling according respective policy. QoS based grid process preparation resource scheduling model and algorithm are proposed. In the proposed algorithm, economic theory is applied to adjust and determine the price and demand of manufacturing grid resources for maximizing the utilities of resource providers and consumers.Fifthly, resource reservation mechanism based on prediction in PPSG is studied. Reservation request, resource performance and time characteristics of tasks are proposed to be considered in resource reservation request. An improved residual error GM (1,1) model is adopted to predict the resource performance and furthermore prediction of time characteristics of tasks are studied with combination of knowledge reasoning. An algorithm of grid resource reservation, acceptance and response is proposed to guarantee the Quality of Service of grid service and simulation experiments are carried out to validate its feasibility and effectiveness.Finally, on the basis of above research, a resource management system of PPSG is developed. The implementation environment and realization method are stated.