General Parametrization Framework

The use of search and optimization algorithms has been increasingly applied on optimization problems in Software Engineering. Several algorithms has been proposed and used in academy and industry. However, the performance of these algorithms is often related to a good configuration of its parameters according to its using context. Aiming to support the identification of the best configuration, this work describes a framework used to optimize the algorithm results obtained with distinct configurations. For that, data of each algorithm configuration tested is collected and stored adequately to be used by statistical tools. Preliminary results indicate the relevance of this framework for the performance optimization of search algorithms.