We study consistent query answering over knowledge bases expressed by existential rules. Specifically, we establish the data complexity of consistent query answering and repair checking under tuple-deletion semantics for a general class of disjunctive existential rules and for several subclasses thereof (acyclic, linear, full, guarded, and sticky). In particular, we identify several cases in which the above problems are tractable or even first-order rewritable, and present new query rewriting techniques that can be the basis for practical inconsistency-tolerant query answering systems.

翻译：我们研究了由存在规则表示的知识库上的一致性查询回答问题。具体而言，针对一类广义的析取存在规则及其若干子类（无环、线性、全称、受保护及粘性），我们建立了在元组删除语义下一致性查询回答与修复验证的数据复杂度。特别地，我们识别了上述问题可处理甚至可重写为一阶逻辑的若干情形，并提出了新的查询重写技术，这些技术可作为实用的容错一致性查询回答系统的基础。