An incremental tree seed algorithm for balancing local and global search behaviors in continuous optimization problems

Population-based tree seed algorithm (TSA), a popular optimization algorithm, was used in this study. The purpose of this study is to develop a TSA via improving its exploitation and exploration capability which is the most important element of the algorithm. Accordingly, the study was aimed at increasing the convergence rate and performance level of the algorithm. The population diversity of the algorithm was studied, and the incremental social learning method in the literature was integrated into a TSA. The new algorithm obtained was called “Incremental tree seed algorithm” (ITSA). Four different ITSA methods were obtained by using four different methods. With the new methods obtained, the TSA method was applied to twelve low-dimensional benchmark functions. The success of the proposed methods on functions was extraordinary, and the results were shown in the tables. At the same time, Wilcoxon p-test analysis, sign test and ROC curve analysis of the obtained results were also performed. According to the results of p-test, sign test and ROC curve analysis, the proposed method was found to be successful. It can be concluded that the proposed method is more robust than its original version.