Abstract
The optimization of Software is a central part of commercial software development and generally requires a significant amount of time and costs. Moreover,there is always the risk that after a large initial investment, it may turn out that no
significant optimizations can be found. An automated search for optimizations
would therefore offer substantial economic benefits.
This thesis explores a novel approach of the use of Genetic Improvement and generative AI to optimize Java source code regarding runtime performance. As a
foundation, the state of the art in both research areas is first examined. In addition,
due to the lack of benchmarks in this field, a benchmark dataset is created. This
benchmark is then used for testing the extension of a Genetic Improvement
framework.
Subsequently, the developed tool for automated Java source code optimization is
applied to a large-scale Java project. While promising results were obtained with
the benchmark, the application to the Java project mainly highlighted the limitations of the implemented approach.
| Date of Award | 2025 |
|---|---|
| Original language | German (Austria) |
| Supervisor | Josef Pichler (Supervisor) |