Research On Key Technologies About The Spatial Network Analysis

Spatial analysis is an important character of Geography Information System (GIS). Spatial network analysis is a prominent part of spatial analysis. In recent years, with the construction of digital urban, geographical networks such as transportation, power and so on have been fast developed, which have made the spatial network analysis more and more important. Therefore, the researches to spatial network analysis become more and more deep. However, owing to the special characters, the efficiencies of spatial network algorithm are not applied to practice. Meanwhile, the current spatial network data models also influence the development of spatial network analysis. So, the efficiencies of spatial network analysis algorithms, the new robust spatial data models and the spatial data arrangements have became the main research topics. This paper intends to put forward new thoughts and methods on path algorithms of spatial network analysis, spatial data models and spatial data arrangements through theory and practice. Main research can be generalized as follows:â‘ This paper systemically analyzes various data structures of path algorithm and concludes that quad heap structure is the most optimize data structure in both space complex and time complex. In order to decrease the time complex of path algorithm, the paper employs the cost function to analysis the improvement of path algorithm. This paper indicates that the heuristic function has big effect on path algorithm after analyzing the A star algorithm. Against the shortcoming of A star algorithm which is the reopen of nodes, the paper designs a new quadheap-based Dijkstra algorithm which has high efficiency and heuristic information. Finally, experiments are conducted to prove the validity of our method.â‘¡This paper reviews the three spatial data models-vector model, tessellation model and hybrid model. Then the paper present a feature-based spatial data model to build the GIS spatial network model, considering the effects of planar enforcement of current spatial data model. Finally, the paper employs the object oriented paradigm and hyper-graph to build the feature-based transportation network model.â‘¢The paper analyzes the significance and difficulties of multiple representations.