从万维网获取的数据，以及由开源“NoSQL”数据库系统管理的数据，往往具有半结构化的特点，查询这类数据的标准语言称为SPARQL。半结构化数据凭借着简单、灵活、弹性等非关系型优点已渗透到越来越多行业领域，进而研究SPARQL查询处理的可靠性和高效性等问题变得尤为重要。本项目通过对SPARQL构子的语义特点、查询处理的原理探究、实际查询需求的分析，以及对当前SPARQL查询处理技术不足的总结，研究SPARQL查询处理的可靠性理论及其优化。由复杂性程度从Well-designed类、到Weakly Well-designed类、再到Beyond Well-designed类依次重点研究：(1) 包容关系的判定边界，包括极大可判定、极小不可判定子类；(2) 限定、低价包容存在关系的复杂度分析；(3) 基于包容存在的优化改写与可靠计算。研究成果将为SPARQL查询处理提供新的可靠性理论和优化方法。

Data obtained from the World Wide Web, as well as those managed by the open-source "NoSQL" database system, are often semi-structured, and the standard language for querying such data is called SPARQL. Semi-structured data are widely applied in more and more fields due to its non-relational advantages such as conciseness, flexibility, elasticity and so on, it becomes important to study the reliability and efficiency of SPARQL query processing. In this project, we study query containment-based theoretical reliability and optimization of SPARQL query processing by analyzing the characteristics of SPARQL operators, the principle of SPARQL query processing, the requirements of SPARQL queries in a real world, as well as the disadvantages of current SPARQL query processing approaches. In this paper, from Well-designed queries, to Weakly Well-designed queries, and to Beyond Well-designed queries, we focus on (1) Decidable boundary of SPARQL query containment including in maximally decidable fragments and minimally undecidable fragments; (2) Computation complexity of existential constrained query containment problems and existential low-level query containment problems; and (3) Optimization based on query rewriting via query containment and reliable query computing. The research results of this proposed project will enrich theoretical reliability and provide more new optimization methods for SPARQL query processing.