Artficial Neural Networks (ANN) are inspired by the structure and functional aspects in biological neural networks. They are trained through mechanisms obtained from the physical properties of the processes involved, for example, electromagnetic waves. From the knowledge acquired through that experience and learning, they may be able to provide solutions for predicting users behavior and providing, within a region of interest, accurate strategy data for projects and sizing. Those who criticized the application of ANN acquired by nature-inspired algorithms, argued that the problems to be faced were usually without complexities, although the conventional methods that were proposed to solve these same problems were not eficient. Some spurious successes have occurred in certain well-behaved environments, but without_exibility when encountering diverse constraints. Adding to these developments, there is the evolutionary openness of computational tools, which has given extraordinary support for deepening techniques to solve and optimize previously unthought problems. In many optimization issues, the quality of a solution is defined by its performance against several conficting goals. Such coficting objectives cannot be signi_cantly reduced to a single value, for example using a weighted sum or other methodology, but must be considered independently of each other. To achieve accurate solutions with reduced computational costs and shorter processing times, we present the Multi-Objective Evolutionary Algorithms (MOEA), as well as Bioinspired Computation (BIC). Combining the advantages of the classical algorithms, the Metaheuristic Algorithms emerged irreversibly. In many optimization problems, the quality of a solution is defined by its performance in relation to several, coficting objectives. Such conficting goals cannot be sensibly reduced to a single value using a weighted sum or another aggregate function, but rather they must be considered independently from each other. Multi-Objective Evolutionary Algorithms (MOEAs) are a natural answer of this kind of evolution. In this work is presented a hybrid bioinspired optimization technique that associates a General Regression Neural Network _ GRNN with the Multi-Objective Bat Algorithm _ MOBA, for the design and synthesis of the Frequency Selective Surfaces _ FSS, aiming its application in data communication systems by difusion of millimeter waves, speci_cally, in the IEEE 802:15:3c standard. The designed device consists of planar arrangements of metallizations (patches), diamond-shaped, arranged over a RO4003 substrate. The FSS proposed in this study presents an operation with ultra-wide band characteristics, its patch designed to cover the range of 40:0 GHz at 70:0 GHz, i.e., 30:0 GHz bandwidth and 60:0 GHz resonance. The upper and lower cuto_ frequencies, referring to the transmission coe_cients scattering matrix (dB), were obtained at the cuto_ threshold at -10dB, to control the bandwidth of the device.
|
