Font Size: a A A

Research On Query Optimization Algorithms In Embedded Distributed Databases

Posted on:2008-06-02Degree:MasterType:Thesis
Country:ChinaCandidate:J J SongFull Text:PDF
GTID:2178360212994271Subject:Computer system architecture
Abstract/Summary:PDF Full Text Request
Marine information service is of great significance on safeguarding marine rights and interests, exploiting marine resources, warning marine disasters and protecting marine environment. Query optimization, as a key technology on improving the database processing performance, has an important impact on the effective implementation of data utilization and real-time information service in the field of marine information. This thesis is base on the research of the marine environment online monitoring and marine disaster intelligent forecasting system, and study deeply the design and implementation of the query optimization algorithm suitable for the embedded distributed database.This thesis first gives the whole architecture and function classification of the marine environment online monitoring and marine disaster intelligent forecasting system, and focuses on the framework design of the marine monitoring data management subsystem based on embedded technology. In view of the requirement of query processing performance on embedded environment in the system, we analyze in detail the impact of the characteristics of embedded real-time database system on query optimization and the characteristics of traditional query optimization methods. Based on this analysis we combine the greedy algorithm and random search algorithms effectively and introduce the fast solving characteristics of the greedy algorithm to the random search process. This can make the algorithms converge rapidly and extend the search space through iterative sampling, and help the algorithms to be out of local minimum. Moreover, the algorithms can control optimization time by setting termination conditions, which not only fulfil the real-time requirement but also better meet the need of different types of transactions. Simulation results show that the improved algorithms have more advantages than traditional optimization methods in term of running time, space complexity and planning quality compared with traditional optimization algorithms. In term of the complicated multi-join query in the distributed architecture of marine monitoring data management system, this thesis introduces systemically the principles and characteristics of particle swarm optimization (PSO) algorithm, and analyzes the effectiveness of applying PSO to solve the multi-join query optimization problem. The algorithm takes the left-deep tree as search space and adopts ordered string encoding to improve the basic formulas in PSO algorithm. We apply this algorithm to the marine monitoring data experiments and get good results. Efficient data transmission is also an important factor in marine information service. This thesis combines the Client/Server communication model and XML technology, and presents an implementation scheme for real-time marine data transmission. This thesis adopts XML-based marine data exchange standard, which can solve effectively the data exchange between different data formats in the heterogeneous platforms.
Keywords/Search Tags:embedded database, multi-join query, query optimization, particle swarm optimization algorithm
PDF Full Text Request
Related items