A new computational intelligence-based metaheuristic algorithm: democracy optimization

Democracy Optimization, as a new Computational Intelligence-based metaheuristic algorithm and optimization method focused on reliability and robustness, is presented along with its conceptual and mathematical designs and the implemented software tool. The designed algorithm introduced a permanent horizontal diversification for the movements of searching agents via the character features defined for the agents. Also, a temporary vertical diversification has been introduced to the movements of searching agents via the hierarchical role features assigned to the agents. With the inclusion of both features in the design, for the first time in literature, the movements of identical search agents have been diversified in a dynamic matrix order. Via this innovation introduced by the algorithm, a balance between the exploration and exploitation functions is set, and the accuracy and reliability of the method were improved. For the algorithm, a certain level of randomness was introduced into the search mechanisms with probabilistic and conditional probabilistic decisions, and thus, contributed to the durability of the method. The optimization software tool is created with the soft implementation of the algorithm. To the realization of the method, the appropriate functioning of the software tool was verified first. Then, it was tested 100 times for each of the selected 16 benchmark functions. Consequently, the optimization performances were measured and then compared with a well-known method. With the evaluation of the obtained performances, the validity and adequacy of the method were demonstrated after showing that it works with high reliability and durability in line with its design focus. The method performed a hyperparameter tuning in a financial market forecasting study, and showed its usability in real optimization problems, and the method was realized. Thus, with its design that includes original search mechanisms diversified in matrix order, Democracy Optimization method and algorithm have made a significant contribution to Computational Intelligence and optimization literatures.