A novel generalized normal distribution arithmetic optimization algorithm for global optimization and data clustering problems

In the recent years of the progress of data volumes and the emergence of complexity in solving large and complex problems, the need for advanced and intelligent methods to deal with these problems has become a necessity. In this paper, we devised an innovative method using a set of elements of intelligent optimization algorithms to solve a set of problems requiring advanced methods to deal with them. In the proposed method, which is called GNDAOA, three main components are used: Arithmetic Optimization Algorithm (AOA), Generalized Normal Distribution Optimization (GNF), and Opposition-based Learning strategy (OBL). These components are used based on a novel transition mechanism to arrange the executions of the used methods during the optimization process to tackle the main weaknesses of the original methods. Two main problems are used to validate the performance of the proposed method; 23 benchmark functions and 8 data clustering problems. The results of the proposed method are compared with several other well-established methods. The proposed GNDAOA method got the best results in 93% of the tested cases of the benchmark functions. It performed very well by a promising behavior to deal with data clustering applications and gained more than 90% improvements compared to the original methods.