Optimal Index Selection Using Optimized Deep Q-Learning Algorithm for NoSQL Database

The resource requirements for managing the data for sophisticated applications have increased as big data technology has advanced. NoSQL (MongoDB) databases are being used increasingly frequently as a result of the desire for high-performance reading and writing. However, the performance of the database is degraded due to the number of queries in a certain time period. Thus to enhance the database performance, an automatic index selection scheme is presented in this paper. Namely, an optimized deep Q-learning network (DQN) is presented for optimal index selection. To enhance the decision making performance of DQN, an adaptive crocodile optimization algorithm (ACOA) is used. Using this algorithm, best action sequences of DQN are obtained. In terms of YCSB mongoDB, the suggested model's performance is assessed. The article's findings show that the suggested model achieves better average cost time (ACT), average time of query execution (ATQ) and query per hour (QPHH) and throughput. © The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd. 2024.