Abstract: As for the problems that server and network have too much pressure and response time is too long during the data transmission of tile, cache mechanism has been established in client side, cache index of tile has been constructed and a tile cache replacement method named Geographic Unit Heat（GUH） has been proposed. The algorithm is based on the smallest geographic unit of tile, which integrates the zoom level of tile and the heat of geographic location, and utilizes heat volatile to adapt to the temporality of tile cache. On the basis of this method, the heat value and cache space have been abstracted into 0/1 knapsack problem and the solution is solved by ant colony algorithm. The results of experiment indicate that this algorithm has significant increase in tile hit rate and byte hit rate.

Key words: cache policy, geographic unit, zoom level, heat, tile

摘要： 针对瓦片数据传输时造成的服务器与网络压力过大、响应时间长的问题，在客户端建立缓存机制，构建瓦片缓存索引，提出了瓦片数据的GUH缓存替换算法。该算法基于瓦片最小地理单元，融合了瓦片的缩放层级和地理位置热度，并利用热度挥发适应瓦片缓存的时间性。在此基础上，将热度价值与缓存空间抽象为0/1背包问题，并通过蚁群算法进行求解。实验结果表明，该算法在瓦片命中率和字节命中率方面有显著提高。

关键词: 缓存策略, 地理单元, 缩放层级, 热度, 瓦片