Research On Campus Electronic Map Based On An Improved Cache Replacement Algorithm

With the development of information technology, many universities have established varieties of campus property management systems. However, they mainly focus on storing and displaying the basic data. In this mode, users cannot get the actual location of a building in a dynamic way and lack intuitive understanding for the overall layout of the campus. This reduces work efficiency of property managers. It shows that we need an interactive visualization platform to improve our management mode.In this paper, we establish a campus electronic map, which is based on public tile map service. Our platform combines three-dimensional modeling techniques, tile cache service, load balance technology and related Web Geographic Information Systems development frameworks successfully.Firstly, we build the basic campus space data model, including main buildings, green space, roads, water systems and other models. Meanwhile, geographic elements are divided into different layers according to their geographic features. And we get the background layer, road layer and point of interest layer; Secondly, we describe the storage model for the map tiles. The original map image is cut into many 256*256 map tiles according to the map zoom level. Thus we get a tile pyramid to store the tiles. Besides, a quadtree is used to index map tiles to improve search efficiency.It is very important to improve the response speed for the campus electronic map, so that our system can be more available and friendly. We propose two methods to optimize the system performance. On the one hand, we establish a map tile caching mechanism, which builds a tile cache pool to store the frequently accessed tiles. If users query tiles in the cache pool, the target tile will be returned directly. What’s more, we introduce an improved cache replacement algorithm. Multiple experiments show that our method outperforms the existing cache replacement algorithms such as Least Frequently Used Algorithm and Size algorithm and presents visually pleasing results with both higher cache hit rate and better robustness. On the other hand, we adopt the consistent hashing algorithm to achieve load balance. Experiments show that this method can distribute user requests to each node evenly. And in the case of removing damaged nodes or create new nodes, our load balancing solution can relocate requests with little cost.In this paper, we propose a possible architecture for a Web platform to support the GIS of Shandong University, which provides users with an interactive and extendable campus map for self-guided tour. By distributing geospatial information and roads to operate it across the Web, a universal broadcast service has been obtained, allowing users to access information about location sensitive data and various services. Furthermore, users can also dynamically acquire multiple details of building interior’s including room amount, room status and even structure blueprints etc. The architecture is designed for Shandong University campus but it is built and developed in such a way that can be conveniently extended to accommodate other interactive maps such as those found in other campuses or parks. We have integrated the campus map into the university property management system. Large numbers of tests demonstrate that our system obtains a satisfactory result thanks to its positive user experiences and great stability.