Framework selection for developing optimization algorithms: assessing preferences by conjoint analysis and best-worst method

In recent years, the evolutionary algorithms used in the solution of NP-Hard problems have become increasingly important. In addition, platforms and application development languages have diversified and started to be differentiated according to their intended use. However, the selection of an appropriate model development environment has become an important decision problem. This study guides the selection of suitable tools for optimization problems, especially in management science. The main objective is to identify the key attributes of the frameworks from the researcher's point of view in management science and assign a total utility score to measure the relative importance of frameworks for evolutionary algorithms. For that reason, we propose a conjoint analysis model upon the preferences of management scientist for the appropriate framework that meets the needs in optimization problems. We also aim at providing effective usage of relevant frameworks for appropriate types of problems, facilitating the work of researchers and therefore increasing the quality of the optimization procedure. By doing so, losing time and effort resulting from the wrong platform and framework selection, as well as ineffective model results, will be avoided. Moreover, the frameworks are also evaluated by calculating the weights of criteria with one of the recent multi-criteria decision-making method called Euclidean best-worst method and compared with the findings obtained from conjoint analysis. This study not only provides review of existing software tools developed for optimization problems but also contributes to research and practice in the field of optimization algorithms in general and helps the researchers in management science for meeting their needs while searching for the appropriate framework