A Dual-Store Structure for Knowledge Graphs

To effectively manage increasing knowledge graphs in various domains, a hot research topic, graph storage management, has emerged. Existing methods are classified to relational stores and native stores. Relational able store large-scale convenient updating knowledge, but the query performance weakens obviously when selectivity of is large. Native efficient processing complex queries due its index-free adjacent property, they inapplicable limited budgets or inflexible process. Motivated by this, we propose dual-store structure which leverages accelerate process store. However, it challenging determine what data transfer from at time. address this problem, formulate as Markov Decision Process derive physical design tuner DOTIL based on reinforcement learning. With DOTIL, adaptive dynamic changing workloads. Experimental results real demonstrate that our proposed improves up average 43.72% compared with most commonly used