基于直方图的空间查询代价评估模型研究及实现

Spatial query cost evaluation is a basic theory of query optimization technique. Nowadays there are two problems in spatial query optimization field. One is that spatial query cost estimation model is not mature; the other is that histograms can only support selectivity evaluation of simple spatial queries. This project will focus on the up-deducing of spatial histograms, the selectivity estimation of each query step and the I/O and CPU cost estimation of spatial query operator. And then, it will form a series of spatial query cost estimation model with higher estimation precision, which should be widely used and easily integrated with database core. The project will also design and implement these models under the support of self-copyright spatial database management system. This work will strengthen research of spatial selectivity estimation in GIS fundamental theory and enhance the background query capacity of huge geographic spatial database.

空间查询代价评估模型是空间查询优化技术的核心基础理论。针对目前空间查询代价评估模型研究不成熟、直方图仅支持简单空间查询选择率估计等现状，项目重点突破空间直方图的向上推演以及查询各环节选择率估计等问题，突破基于I/O、CPU的空间查询代价评估模型，形成一套适用范围较广、估计精度较高、易于与数据库内核集成的空间查询代价评估模型，并在自主版权空间数据库管理系统（BeyonDB）的支持下设计实现，以期加强GIS基础理论中空间选择率估计的研究，提升海量地理空间数据库的后台查询承载力。

空间查询代价评估模型是空间查询优化技术的核心基础理论。针对目前空间查询代价评估模型研究不成熟、直方图仅支持简单空间查询选择率估计等现状，项目重点突破空间直方图的向上推演以及查询各环节选择率估计等问题，突破基于I/O、CPU 的空间查询代价评估模型，形成一套适用范围较广、估计精度较高、易于与数据库内核集成的空间查询代价评估模型，并在自主版权空间数据库管理系统（BeyonDB）的支持下设计实现，以期加强GIS 基础理论中空间选择率估计的研究，提升海量地理空间数据库的后台查询承载力。