| Since the isomerism, diversity, distribution and management multiplicity of manufacturing resources, it gives the manufacturing grid (MGrid) resource management a great challenge. In order to solve these problems, many researchers have studied resource encapsulation, information services, quality of service (QoS), and optimal allocation etc. However, there is few systematic theory and methodology concentrating on economic phenomena and resource value in the MGrid resource management. The insufficient research mainly reflects in the following aspects:(1) How to trade and make a price in the MGrid system, meanwhile guarantee the authenticity of resource information? (2) How to semantically describe resources for users in order to publish resource information? (3)How to realize the empirical search in order to avoid traversing all resource information? (4)How to realize the resource reservation efficiently, especially when there are large amounts of co-reservation? (5)How to schedule resources with guaranteeing MGrid QoS?This doctoral dissertation focuses on the above questions. Based on the in-depth research on the MGrid resource management, the dissertation proposes a MGrid resource management architecture (MGridRMA). According to the MGridRMA, a platform of MGrid resource market is designed to support multiple trading models. The possible economic phenomena in the MGrid resource market are systematically analyzed. The system solutions of MGrid resource management based on economics are given. Finally, the experimental platform is developed in order to verify the proposed theory and methodologies. The main contributions and works of this dissertation are as follows:(1)The principles of economics involved in the MGrid resource management are analyzed. The characteristics of MGrid resource market are given out and the model of MGrid resource market is established. The technology stack of MGrid resource market is designed, in order to demonstrate the important factors for establishing a MGrid resource management system based on economics. The framework of MGrid resource management based on economics is realized, utilizing the grid core middleware. The economic relationships between all entities and the information asymmetry in the MGrid resource market are analyzed. Four kinds of common resource trading models are built on the MGrid resource market and then the flow of resource management is designed.(2)The credit mechanism based on perfect Bayesian equilibrium is proposed. Then model the mechanism, solve and analyze the model. Finally, take a simulation experiment on the mechanism and analyze the experimental data.(3)The mechanism of MGrid resource reservation is proposed. The reservation life cycle is studied. The model of is classified into three categories. The architecture of MGrid resource co-reservation is designed to support the two models of reservation. A negotiation flow and a series of application programming interfaces (API) of co-reservation are given out. The processing of co-reservation requests is transferred to the subgraph isomorphism problem and an improved algorithm is introduced to solve it efficiently.(4)Based on the quantum evolution theory, a hybrid chaotic quantum evolutionary algorithm (CQEA) for resource scheduling is proposed. The requirements of QoS in the MGrid system are given out. A layer model of MGrid QoS is designed. The CQEA is used to meet MGrid QoS in the process of resource scheduling. The experimental results show that the proposed CQEA is effective, efficient and scalability.(5)The auto parts oriented MGrid resource management system (APOMGridRMS) is developed and implemented. The APOMGridRMS tries to implement resource search, production procurement, resource reservation and scheduling and so on. The main object of APOMGridRMS is to run and purchase all kings of auto parts acting as an agency for all its clients from all over the world. The APOMGridRMS helps us make the proposed theory perfect and revise the methodologies.
|
PDF Full Text Request