Year: 2025
Type: article
Abstract: Today, as engineering problems become more complex in terms of the effective variables in these problems and the range of their changes and their multidimensionality (in terms of not being able to look at a problem from one dimension and its various dimensions must be considered) And the need to trade off on the goals of these issues, especially issues that are in conflict with each other and the need for simultaneous optimization of these goals, the need to use multi-objective optimization methods has become more apparent. For example, in the design of VLSI circuits that have both power and latency, given that these two parameters have opposite behaviors, that is, by increasing one of them, the other decreases, and we need to optimize both of these parameters together. In such problems, using multi-objective optimization methods, the most optimal state for these two parameters can be obtained simultaneously. Since one-objective optimization methods have shown their ability to deal with various engineering problems, so often with the introduction of any one-objective optimization method after a while, the multi-objective optimization method Based on the same method has been presented by researchers to the scientific community to test and evaluate the ability of optimization based on the same single-objective method in the field of multi-objective optimization. For example, the NSGAII multi-objective optimization algorithm is based on the GA single-objective optimization algorithm, the MOPSO algorithm is based on the PSO, the MOGSA algorithm is based on the GSA algorithm, and the MOIPO algorithm is based on the IPO algorithm. Today, more than 350 one-objective methods have been introduced to the scientific community, which are widely used in various researches and examined and tested, and in terms of quantity and quality, the development of these one-purpose methods is a competition field Researchers have provided, and no necessary effort has been made in multi-objective versions, and as many as the single-objective method has been introduced to the scientific community, the multi-objective method has not been introduced based on those methods. Therefore, in this dissertation, we have tried to multi-objective this method by considering a recently introduced one-objective optimization method called Arithmetic Optimization Algorithm (AOA). Add existing multi-objective methods and enable researchers to use the ability of this single-objective method to deal with multi-objective problems. In this regard, in order to multiply the goals of the arithmetic optimization algorithm, the concept of Pareto optimality has been used to detect non dominated solutions and repository to store these solutions. To measure the performance of the Multi-objective Arithmetic Optimization Algorithm (MOAOA), this algorithm is applied to the famous benchmark functions and to compare the performance of this algorithm with the famous NSGAII and MOPSO algorithms, spacing and generational distance criteria have been used. Based on the results of these experiments, it was found that this algorithm shows acceptable performance compared to popular optimization algorithms.
Author: Amel Nashat Shakir
Institution: Northern Technical University
Language: English
Subfield: Computational Theory and Mathematics
Field: Computer Science
Domain: Physical Sciences
Open Access status: diamond
OpenAlex ID: W4415605919