| Popis: |
Existing optimization algorithms are insufficient in the face of problems with difficult measurement functions as well as a large number of design parameters. Therefore, to achieve such challenging applications, the Chameleon swarm algorithm and its variants, which belong to the family of meta-heuristic algorithms that have not been used in any antenna optimization study before, are used together with 3 different objective functions fitted from mathematical models. Then, a U-slot antenna application with 12 different variable design parameters with resonance frequencies of 3.5 GHz, 3.7 GHz, 5.2 GHz and 5.8 GHz is considered as a multidimensional and single-objective optimization problem. In this study, first of all, the success of the algorithm is reinforced by comparing the performance with other commonly used single and multi-objective optimization algorithms. In addition, the results obtained with different population parameters, weight coefficients, objective functions and variant models were compared. All these processes are compared within themselves, and the antenna results of the most successful result are displayed as 3D electromagnetic simulation. The results show that the optimization processes proposed for an antenna designer are a safe, practical and efficient solution for multidimensional and single-objective antenna optimization applications. In addition, any optimization problem with a large number of variable design parameters can undoubtedly be adapted to the Chameleon swarm algorithm. |
