ABSTRACT
Query optimisation is among the most essential and very well problems in database systems. Traditional query optimisers, on the other hand, are complicated heuristicdriven algorithms that take a long time to tune for a specific database and much more moments to construct and preserve in the first place. By weighing the various query plans, the query optimiser tries to figure out the most efficient way to execute a query. Attempts utilise reinforcement learning techniques to query optimisation problems have shown promise. In this research article, researchers explore how a variety of query optimisers based on deep reinforcement learning can vastly outperform the current state-of-the-art. Researchers spot potential stumbling blocks for future research that combines deep learning with query optimisation, and also unique deep learning-based methods that might lay the foundations for learning-based quick and appropriate from start to finish.
