Desenvolvimento de extensão de VS Code para a análise de complexidade de algoritmos

Abstract

Understanding how algorithms behave in terms of execution time is essential for optimizing software performance. However, the theoretical analysis of complexity is not easy or directly applicable in practice, as developers face challenges in estimating the efficiency of their code without adequate tools. To address this need, an extension for Visual Studio Code was implemented that analyzes the complexity of algorithms, integrating directly into the software development process. This work focused on developing a practical solution after a theoretical review and evaluation of existing solutions. The results obtained showed that the extension can analyze basic control flow structures using AST and provide more accurate results through natural language models, delivering correct asymptotic estimates. The automated analysis turned out to be effective, bridging the gap between theory and practice, and allowing developers and students to quickly assess the complexity of their algorithms, achieving around 96% accuracy. The validation of the developed solution confirmed that the integrated analysis into the development workflow is not only feasible but also beneficial, promoting greater awareness of computational complexity and encouraging good coding practices, thus facilitating the practical application of theoretical concepts of complexity. [resumo fornecido pelo autor]