Thesis:

Optimization of Join Queries in Distributed Databases Using the Orchard Algorithm

This research addresses the issue of optimizing the execution of join queries in distributed databases. With the increase in data volume and the need for concurrent processing, optimal performance in such databases becomes critical, as join queries are among the most costly operations due to large data exchanges across nodes. The objective of this study is to propose an efficient method based on the Orchard Algorithm, inspired by natural plant growth processes, to create a balanced combination of global search and local improvement. The methodology combines conceptual modeling, algorithm simulation in the MATLAB environment, and real implementation in the PostgreSQL database. During this process, a population of query plans is generated, evaluated, and optimized through stages of growth, competition, and update. Test data from the standard TPC-H suite was used for performance evaluation. Results showed that compared to other algorithms such as Harmony Search (HS) and Artificial Bee Colony (ABC), the Orchard Algorithm performed better on key metrics including reduction of Query Evaluation Cost (QEC), execution time (Execution Time Function (ETF)), and system resource usage (Resource Usage Function (RUF)). Additionally, the Query Plan Quality Index (QPQI) improved by approximately 10%, and the convergence rate was up to 30% faster than competitors. These findings indicate that the Orchard Algorithm is an effective option for optimizing complex queries in distributed database systems.

Keywords: Keywords: query optimization, join queries, distributed database, Orchard Algorithm, metaheuristic a