Composition of Static and Dynamic Analysis for Algorithmic-Level Code Semantic Optimization

Program optimization can be done at various levels like design, algorithm, data structure, source code, and run time. It can be based on various parameters like time, space, and code complexity. Higher the level of optimization, the more significant the impact. This paper proposes algorithm-level optimization. Programs with similar semantics are analyzed on a composite metric dependent on factors such as time, space, cyclomatic complexity, and code quality of snippets to suggest the most optimal algorithmic solution. Factors like time, space are calculated dynamically and approximated using polynomial regression. Factors like cyclomatic complexity and Halstead difficulty metric are calculated statically.