基于图嵌入的图匹配算法分析-光学工程专业论文.docxVIP

西安建筑科技大学硕士学位论文 西安建筑科技大学硕士学位论文 西安建筑科技大学硕士学位论文 西安建筑科技大学硕士学位论文 Reach on graph matching algorithms based on the graph embedded Specialty: Signal and information processing Name: Pengde Yue Instructor: Prof. Changhua Li ABSTRACT As the social progress and the rapid development of science and technology, the information in the Internet age has skyrocketed. Computer data are faced with massive data processing and data diversity enormous challenges. Data expressed in Fig data structures, more data is playing the role. Map data is discrete mathematics and computer science important data structure in which graph matching problem in many areas have been greatly attention. Currently the graph matching method is applied to the problem of building information modeling retrieval few studies, architectural space is connected into the form of performance graph topology. However, graph matching problem is a NP-hard problem, people made many based on statistical distribution, pattern recognition, the theory of intelligent optimization algorithms to solve the problem. Overall, the problem is not completely solved, Individually advantages and disadvantages of each method. In this paper, We are research the relationship between some of the major algorithms and their characteristics of existed. In the analysis of the existing graph embedding method for graph matching problem, based on the vector space exploration for orthogonal graph embedding algorithm, to establish the best set of orthogonal vector space method, so that the original algorithm is effective and applicable is improved. Improved algorithm combined with multi-dimensional vector space coordinate system established set of ideas that will map to quantify, orthogonal and quantitative expression vector space embedded into focus, quickly get the most effective match combinations. In the graph nodes corresponding matching problem in the optimization process for retrieval and other issues facing the combinatorial explosion, the pap