Search-based refactoring: an empirical study

Object-oriented systems that undergo repeated addition of functionality commonly suffer a loss of quality in their underlying design. This problem must often be remedied in a costly refactoring phase before further maintenance programming can take place. Recently search-based approaches to automating the task of software refactoring, based on the concept of treating object-oriented design as a combinatorial optimisation problem, have been proposed. However, because search-based refactoring is a novel approach it has yet to be established which search techniques are most suitable for the task. In this article we report the results of an empirical comparison of simulated annealing, genetic algorithms and multiple ascent hill-climbing in search-based refactoring. A prototype automated refactoring tool is employed, capable of making radical changes to the design of an existing program in order that it conform more closely to a contemporary quality model. Results show multiple-ascent hill climbing to outperform both simulated annealing and genetic algorithm over a set of five input programs. Copyright c © 2000 John Wiley & Sons, Ltd.