Join Query Optimization with Deep Reinforcement Learning Algorithms
This work addresses performance bottlenecks in database query processing for data-intensive applications, representing an incremental improvement by applying existing DRL techniques to a known problem.
The paper tackles the NP-hard problem of join query optimization by introducing a deep reinforcement learning (DRL)-based framework called FOOP, which avoids exhaustive enumeration and is significantly faster than traditional dynamic programming methods, with Proximal Policy Optimization outperforming Q-learning algorithms.
Join query optimization is a complex task and is central to the performance of query processing. In fact it belongs to the class of NP-hard problems. Traditional query optimizers use dynamic programming (DP) methods combined with a set of rules and restrictions to avoid exhaustive enumeration of all possible join orders. However, DP methods are very resource intensive. Moreover, given simplifying assumptions of attribute independence, traditional query optimizers rely on erroneous cost estimations, which can lead to suboptimal query plans. Recent success of deep reinforcement learning (DRL) creates new opportunities for the field of query optimization to tackle the above-mentioned problems. In this paper, we present our DRL-based Fully Observed Optimizer (FOOP) which is a generic query optimization framework that enables plugging in different machine learning algorithms. The main idea of FOOP is to use a data-adaptive learning query optimizer that avoids exhaustive enumerations of join orders and is thus significantly faster than traditional approaches based on dynamic programming. In particular, we evaluate various DRL-algorithms and show that Proximal Policy Optimization significantly outperforms Q-learning based algorithms. Finally we demonstrate how ensemble learning techniques combined with DRL can further improve the query optimizer.