Machine Learning in Antenna Design: An Overviewon Machine Learning Concept and Algorithms

Hilal M. El Misilmani and Tarek NaousElectrical & Computer Eng. Department

Beirut Arab UniversityBeirut, Lebanon

hilal.elmisilmani@ieee.org, tareknaous@ieee.org

Abstract—With the growth and wide variety of available data,advanced processing, and affordable data storage, machine learn-ing is witnessing great attention in finding optimized solutions invarious fields. Machine learning techniques are currently takinga major part of the ongoing research, and expected to be thekey player in today’s technologies. This paper introduces andinvestigates the applications of machine learning in antennadesign. It covers the major aspects of machine learning, includingits basic concept, differentiation with artificial intelligence anddeep learning, learning algorithms, its wide applications invarious technologies, with a main focus on its usage in antennadesign. The review also includes a comparison of the resultsusing machine learning in antenna design, compared to theconventional design methods.

Index Terms—Machine learning, artificial intelligence, antennadesign, neural networks, electromagnetic simulations

I. INTRODUCTION

Artificial Intelligence (AI) is the art of enabling machinesto perform tasks that require human thinking abilities, suchas learning, decision making, and problem solving. In sim-ple words, AI is implementing human thinking ability inmachines. With the latest advances in big data availability,software engineering capability and affordable high computingpower, AI is becoming an essential part of today’s research.It is expected to affect most of our daily activities, with fun-damental changes in science and engineering, and enormousimpacts on society, with the ability to create, transform, oroptimize different aspects and applications of our daily lives.

To be able to implement and achieve artificial intelligence,several capabilities need to be possessed by the computer ormachine. For instance, natural language processing is requiredfor successful communication in English and knowledge rep-resentation is needed for storing information. To answer ques-tions and drawing conclusions using the information stored,automated reasoning is required, whereas, to derive patternsin data and make predictions and adapt to new conditions,machine learning is necessary. For object perception andmanipulation computer vision and robotics are needed. Byusing intelligent algorithms and large amount of data sets,iterative processing allows AI software to learn automaticallyfrom patterns or features [1].

Artificial Intelligence and Machine Learning (ML) are oftenused interchangeably, however, as will be discussed hereafter

Fig. 1: Relationship between Artificial Intelligence, MachineLearning, and Deep Learning

and investigated in this paper, machine learning is a largesubset of artificial intelligence, as shown in Fig. 1. In fact,machine learning can be considered as an approach to achieveAI applications.

Machine Learning is briefly described as getting usefulinformation out of data, achieved by developing reliable pre-diction algorithms. These algorithms can be very powerful inoptimization, but their success rely on the condition and sizeof the data collected. Therefore, machine learning is frequentlyassociated with statistics and data analysis [2], [3].

As for Neural Networks, they are defined as a type ofmachine learning algorithms that tries to imitate how thehuman brain works. They consist of layers of interconnectednodes. Each node produces a nonlinear function of its input.Deep Neural Networks (DNNs) are neural networks that havemore than one hidden layer [4], usually refereed to as DeepLearning. Both of these algorithms are considered as types ofmachine learning algorithms.

This paper presents an overview on machine learning, with amajor focus on investigating its usage in antenna design appli-cations. The concept of machine learning is studied along withits different learning algorithms. Next, an extensive reviewof several antenna designs and electromagnetic computationalmethods is investigated. The methods used to design theantennas using machine learning in each is presented, alongwith the outcome of each algorithm used.

TABLE I: The Three Main Categories Of Learning

Learning Category Description

Supervised Learning A model is trained on a data setPredictions are made on new inputs

Unsupervised Learning A pattern is derived from the dataafter exploring it

Reinforcement Learning Model takes decisions and learnsfrom its actions

Fig. 2: Schematic of an artificial neural network

II. MACHINE LEARNING OVERVIEW

Machine learning is based on algorithms that can learn fromdata without relying on rules-based programming [5]. It can begenerally divided to three key categories: supervised learning,unsupervised learning, and reinforcement learning [2], [6]–[8],listed in Table I. Additional learning scenarios that are lesswidely used include: Semi-supervised learning, TransductiveInference, On-line learning and Active Learning [2]. Eachcategory has different learning algorithms that belong to it.

In supervised learning, what a correct outputs looks like isalready known. After training the learning algorithm on a givendata set, the algorithm generalizes to give accurate predictionsto all possible inputs [7]. Supervised learning algorithmsinclude but not limited to: linear regression, logistic regression,artificial neural networks, and support vector machines.

– Linear Regression [9]: consists of fitting a continuouslinear function through the data from which the algorithmcan make predictions on new inputs.

– Logistic Regression [9]: used in classification tasks whereit predicts the probability that a certain input correspondsto one of the already known classes.

– Artificial Neural Networks [9], [10]: in neural networks,large interconnections of ”neurons”, which are simplecomputing cells, are employed to achieve good perfor-mance. When complex functions with many features arefound, neural networks offer an alternate way to performML. Neural networks are made of multiple layers, theinput layer, the output layer, and hidden layers betweenthe input and output layers, as shown in Figure 2. One ofthe many types of neural networks is the feed-forwardneural network where a weighted sum of connected-neurons output are received by each neuron as an input.

– Support Vector Machines (SVMs) [10], [11]: another typeof supervised learning algorithm. In particular, it is usedin classification and deals with the more difficult case ofnon-linearly separable patterns by using kernel methods.

– k Nearest Neighbors [9]: considered among the simplestof all machine learning algorithms. After memorizing thetraining set, the output of any new input is predicted bythe algorithm based on the outputs of its closest neighborsin the training set.

In unsupervised learning, what the results should look like isnot known, the algorithm derives a structure from the data afteridentifying similarities in the inputs [7]. Unsupervised learningalgorithms include but not limited to: K-Means Clustering andDimentionality Reduction Algorithms.

– K-Means Clustering [3]: k-unique clusters are automati-cally formed by this algorithm. It is a type of unsuper-vised learning where variables in the data are groupedtogether based on relationships among them.

– Dimentionality Reduction Algorithms [9]: such as thePrinciple Component Analysis Algorithm (PCA) wherethe goal of this algorithm is minimizing the projectionerror by reducing the reducing every feature’s distance toa certain projection line.

In reinforcement learning, no labeled data set is received bythe machine. Instead, information is collected after interactingwith the environment through different actions. The machine isrewarded after each action, hence its objective is maximizingthis expected average reward where the action would becomeoptimal. An example of a reinforcement learning model is theMarkov Decision Process (MDP). [2], [8]

Other machine learning algorithms exist with less widelyusage, such as: Decision Trees [12], Boosting [13], NaiveBayes [9], Bayesian Regularization [14], Kriging [15], andmore [9].

When solving real world machine learning problems, thedata set is split into three parts [9]: a training set used to trainthe algorithm, a cross-validation set used for model selection,and a test set used for testing the performance of the predictionalgorithm, that is checking if it succeeds to generalize oninputs not previously seen.

Machine learning techniques have been widely employed incommunication technology, such as antenna selection in wire-less communications [16], [17], smart grid networks wheremachine learning can detect malicious events before occuring[18], wireless networks where ANNs can be used for predict-ing the mobility patterns and content requests of wireless users[6], speech recognition where SVMs with kernel functionscan be useful to improve its generalization performance [19],context aware computing in IoT [20], and much more [21].

III. MACHINE LEARNING IN ANTENNA DESIGN

Several papers have investigated the applications of ma-chine learning in antenna design. It is expected that machinelearning can provide accelerated antenna design process whilemaintaining high accuracy levels, with a minimization oferror and time saving, along with a possible prediction ofthe antenna behavior, a better computational efficiency andreduced number of necessary simulations.

In general, in order to apply machine learning in antennadesign, the following steps can be done:

Fig. 3: Reflectarray unit cell [26]

1) By multiple simulations, the electromagnetic character-istics of an antenna are found out.

2) These characteristics are stored in a database and usedas a data set for training a certain machine learningalgorithm.

3) The antenna that gives the closest results is designed bythe algorithm after making predictions, depending on theneeds of the designer.

In this section, a detailed review of several antenna de-sign papers using machine learning is provided. For a clearinvestigation, the papers discussed here are divided to twomain categories: machine learning for parameter optimization,and machine learning for enhancing evolutionary computationalgorithms.

A. Machine Learning For Parameter Optimization

One approach of using machine learning in antenna designis training a learning algorithm on data collected from previoussimulations to optimize the antenna parameters.

Based on Local-Periodicity (FW-LP), successful analysisand optimization has already been achieved using a Full-Waveanalysis [22], [23], however, for space applications, fastercomputations are required. Although this has been addressedpreviously using Artificial Neural Networks (ANNs) [24],[25], the results were limited. In [26], the design of shaped-beam reflectarray unit cell using Support Vector Machine(SVM) has been presented. Two sets of four parallel dipolesare used in the design, shown in Fig. 3. To accelerate the reflec-tarray antennas design machine learning algorithm is used. Thereflection coefficient matrix is first characterized using SVM.The scattering parameters of the unit cell dimensions are thenderived. The influence of the discretization of the angle ofincidence is also taken into account in the simulations. Theresults in paper using SVM are then compared to those in[27] based on the Method of Moments (MoM-LP) in termsof speed and accuracy. The results showed that the sequentialdesign using SVM can be accelerated by a factor of 880, andthe parallel design can also by accelerated by a factor of 566.Fig. 4 showing the gain pattern comparison of the SVM andthe MoM-LP using the real angles of incidence.

(a)

(b)

Fig. 4: Gain pattern comparison of SVM and MoM (a)elevation and (b) azimuth [26]

The design of a rectangular microstrip antenna using Sup-port Vector Machines (SVMs) has been also presented in [28].SVMs are used here for regression, and hence it may be calledSupport Vector Regressors (SVR). The machine learning al-gorithm is trained on a dataset that includes measured valuesof the input impedance, operation bandwidth and resonantfrequency of the antenna. Artificial Neural Networks are alsoemployed at the same time to compare the results with thoseobtained using SVR. It was shown that employing machinelearning algorithms in the design of microstrip antennas givesbetter and more accurate characterization than the theoreticalresults. Comparing SVR and ANN, it was shown that SVRprovides better computation efficiency since it has a fasterconvergence rate, with less training time and test time neededin comparison to ANN.

The design of a planar inverted-F antenna (PIFA), shownin Fig. 5, with magneto dielectric Nano-composite (MDNC),using the Bayesian Regularization algorithm as the neuralnetwork learning process, has been presented in [14]. Startingwith the nano-magnetic material’s volume fraction and particleradius, the different antenna parameters, such as radiationefficiency, gain, resonant frequency, bandwidth and others, can

(a)

(b)

Fig. 5: PIFA antenna structure: (a) top view, (b) side view [29]

Fig. 6: Machine learning output vs target [29]

be calculated with high accuracy using machine learning. Forthis design two databases are created. The first one containsmaterial properties (volume fraction and particle radius). Thesecond one contains electrical properties of the material,such as electric and magnetic loss tangent, permittivity andpermeability. To create a relation between the performanceof the antenna and its properties, machine learning algorithmis trained using the two databases as input and output. Newmodels for permeability and permittivity of the compositematerial are also proposed in [14]. Using these equations withonly 42 samples created, a total error of only 7% is obtainedbetween machine learning data and the second database. Fig.6 shows a comparison of the Machine Learning Output andthe target (all of the data stored in second database) regressioncurve. It was concluded that machine learning techniques canbe very useful in minimization of error and acceleration of thecycle time for new materials synthetization, and in expectingthe behavior of the antenna without performing extensive timeconsuming simulations. Optimized results can be obtained byincreasing the number of data samples used with the algorithm.

Fig. 7: A comparison of the accuracy and time saving usingNeural Networks [15]

The design of reflectarrays using the Kriging algorithminstead of the classical full-wave solvers has been presented in[15]. Machine learning is used here to predict the response ofcomplex unit cells to design high-performance reflectarrays.The scattering matrix of complex reflectarray elements ispredicted while reducing computational time. The model istrained on a training set of N samples, where N ∈ [500,20000],to learn the input output relationship. A comparison of theaccuracy and time saving using the proposed method is shownin Fig. 7. It was concluded that using the Kriging model, a 99%time saving can be reached, while maintaining a predictionerror below 5%.

The design of a W-Band slotted waveguide antenna array,shown in Fig. 8, by employing artificial neural networkshas been also presented in [30]. The input layer of theneural network consists of seven design parameters whichare the lengths and orientation angles of the coupling slots(lc, θ1, θ2, θ3, θ4, θ5), in addition to the length of the radiatingslots (lR). A data set of 189 examples are obtained by simu-lations in HFSS and stored in a database which will be usedfor training, cross-validation and testing of the model. Afterobtaining the optimized values of the design parameters, theantenna is fabricated using the SLA 3D printing techniques.The simulated S11 characteristic as well as the gain of theantenna are compared to those measured and a good agreementwith some slight errors has been shown.

B. Machine Learning For Enhancing Evolutionary Computa-tion Algorithms

Another approach of using machine learning in antenna de-sign is embedding it in Evolutionary Algorithms. EvolutionaryAlgorithms (EAs) are taking important interest in computingsciences. Using evolutionary mechanisms involved in naturalselection, evolutionary algorithms allow to find approximatesolutions to optimization problems. EAs have been used inprevious works [31]–[33] for the purpose of antenna designand optimization. To improve their performance, machinelearning techniques have been used successfully in Evolution-ary algorithms. When using EAs, databases of adequate size

(a)

(b)

Fig. 8: Slotted Waveguide Antenna (a) top view, (b) bottomview. [30]

are produced which allows to use machine learning techniques[34].

Particle Swarm Optimization Algorithm (PSOA), a type ofevolutionary algorithm, is used in [35] to design a multi-band patch antenna using artificial neural networks. After thegeometrical parameters of the antenna are decided by thePSOA, a mapping function is built by the ANN so that thefrequencies and associated bandwidths can be related to thedimensional parameters of the antenna. The flow chart of theoptimizer used with four input parameters, is shown in Fig.10. The results showed that the design process can be speedup by eliminating the need for time-consuming simulations,therefore reducing the computational burden significantly. Thedesign antenna has been also fabricated and tested, with goodanalogy between the measured and simulated results in [35].

In [36], the design of an E-shaped antenna by combiningdifferential evolution (DE) with the Kriging algorithm, hasbeen presented. Six antenna parameters are optimized, the feedposition Px, the slot position Ps, the patch width W, the slotwidth Ws, the patch length L and the slot length Ls, shown inFig. 11. It is suggested that similar results of other optimizationtechniques can be obtained while reducing the number ofnecessary simulations by 80%. The magnitude of the S11

parameter is desired to be minimized at frequencies 5.0GHzand 5.5GHz. After the proposed algorithm was run 5 times,optimum solutions were found, and good prediction accuracywas exhibited by the model. The predicted and simulated S11

Fig. 9: Regression plot [35]

Fig. 10: Flow chart of proposed modified optimizer [35]

curves are also shown in Fig. 12. Combining machine learningwith evolutionary algorithms has been proven to provide afaster convergence rate with a similar solution quality incomparison to optimization methods, self-adaptive differentialevolution [37] and wind driven optimization [38]. It was shownthat the same optimization goals can be reached while reducingthe number of simulations needed by self-adaptive differentialevolution [37] and wind driven optimization [38] by 82.3%and 77.9% respectively.

The design of an ultrawide band microstrip antenna by

TABLE II: Comparison of the different machine learning techniques used in the investigated papers

Paper Antenna Type LearningAlgorithm Used Compared To Results

[26] Reflectarrays SVM MoM & ANN Accelerated design process whilemaintaining high accuracy levels

[14] Planar Inverted F-antenna (PIFA) Bayesian Regularization Minimization of error and acceleration ofcycle time for new materials synthetization

[15] Reflectarrays Kriging Time saving can reach 99.9% whilemaintaining a prediction error below 5%

[29] Planar Inverted F-Antenna (PIFA) ANN Conventional Simulations Possible prediction of antenna behaviorwithout extensive electromagnetic simulations

[28] Rectangular Microstrip Antenna SVM ANN Better computation efficiency witha faster convergence rate

[30] Slotted Waveguide Antenna ANN Conventional SimulationsComputation of several antenna parameters

with good agreement with the simulated andfabricated results

[35] Antenna (SWA) ANN Conventional Simulations Design process can be sped up by eliminatingthe need for time-consuming simulations

[36] Stacked Patch Antenna Kriging Conventional SimulationsSimilar results of other optimization techniquescan be obtained while reducing the number of

necessary simulations by 80%

[39] E-Shaped Antenna Linear Regression Conventional Simulations The optimum results were found without anynecessary simulations

[40] Microstrip Antenna Gaussian Process ML Differential EvolutionThe speed of the design and optimization

procedure by more than four timescompared with differential evolution

Fig. 11: E-shaped patch antenna [36]

combining regression with an evolutionary algorithm has beenpresented in [39]. Machine learning technique is employedhere along with evolutionary algorithm in the estimation offitness function behaviors: the bandwidth (BW), the return loss(RL) and the central frequency division (CFD). The regressivemachine learning algorithm used allows to fit a curve througha discrete set of known data points which are the antennaparameters obtained previously by simulations. A prototypealgorithm has been used to find the optimized values forWs and Ls, shown in Fig. 13, while keeping other antennaparameters constant and considering design restrictions (BW> 9.0GHz, RL < -20dB and CFD < 0.37Hz). It was shownthat, using 170 datasets, the optimum results were foundwhile meeting the required restrictions. The perception on thebehavior of the objectives (BW, RL and CFD) increases as a

Fig. 12: The simulated and predicted S11 curves of the optimaldesign [36]

Fig. 13: Microstrip antenna with investigated parameters [39]

larger number of datasets is used.The design of inter-chip antenna, a four-element linear array

antenna and a two-dimensional array have been presented in

[40]. A new method for designing antennas called surrogatemodel assisted differential evolution for antenna synthesis(SADEA) has been proposed. This method combines GaussianProcess Machine Learning and a Differential Evolution (DE)Algorithm. Using this method in the design of the three differ-ent antennas, the results showed that, using machine learning,SADEA enhanced the speed of the design and optimizationprocedure by more than four times compared with DE. In[41] Less time consumption has been also achieved in thedesign of an isotropic wire antenna by combining a machinelearning technique with an evolutionary algorithm, the fitnessproportionate selection.

By employing Support Vector Machines (SVMs), the designof a regular microstrip antenna has been presented in [42].Better characterization with higher accuracy is achieved byusing SVMs. A comparison of the obtained results with theresults of a different approach that uses Artificial NeuralNetworks (ANNs) is also presented.

A summary of the different papers using machine learningin antenna design are summarized in Table II.

IV. CHALLENGES IN MACHINE LEARNING

Although Machine Learning is very useful, it comes witha lot of challenges. The most common challenges include butare not limited to:

1) The choice of learning algorithm: It is not easy todecide what algorithm to choose as there are a greatnumber of them. This depends directly on what is beingpredicted and also on the type of data acquired. A goodpractice is to always visualize the data before choosingthe algorithm.

2) Problem Formulation: Beginning with wrong assump-tions usually lead to worthless results that can cost alot of time. It is necessary to know what area of theproblem is best to spend time working on.

3) Getting enough data: Some data can be hard to findor obtain. In antenna design, multiple simulations areneeded in order to obtain a training set.

4) Pre-processing of data: To ensure that the learningalgorithm performs adequately, multiple steps need tobe performed on the data, such as data cleaning, nor-malization and feature selection, which would cost timein case of very large datasets.

5) Debugging the algorithm: Knowing what to do next canalso be challenging. When problems, such as high biasand high variance occur, it is crucial to know whatsteps to take. This requires following some diagnosistechniques such as plotting the learning curves.

V. CONCLUSION

This paper presented an overview on machine learning withan investigation of its concepts, differentiation with artificialintelligence and deep learning, its different learning algo-rithms and techniques. An extensive investigation is presentedon the usage of machine learning in antenna design, forwhich its advantages with respect to the traditional design

and computational techniques have been also studied. It wasseen that machine learning can provide accelerated antennadesign process while maintaining high accuracy levels, with aminimization of error and time saving, along with a possibleprediction of the antenna behavior, a better computationalefficiency and reduced number of necessary simulations.

REFERENCES

[1] S. J. Russell and P. Norvig, Artificial Intelligence: A Modern Approach,3rd Edition. Pearson Education, 2016.

[2] M. Mohri, A. Rostamizadeh and A. Talwalkar, Foundations of MachineLearning. The MIT Press, 2012.

[3] P, Harrington, Machine Learning in Action. Manning Publications, 2012.[4] R. Jiang, X. Wang, S. Cao, J. Zhao and X. Li, “Deep Neural Networks

for Channel Estimation in Underwater Acoustic OFDM Systems,” IEEEAccess, vol. 7, pp. 23579–23594, 2019.

[5] D. Pyle and C. S. Jose, “An executive’s guide to machine learning,”tech. rep., McKinsey & Company, June 2015.

[6] M. Chen, U. Challita, W. Saad , C. Yin and M. Debbah, “Machinelearning for wireless networks with artificial intelligence: A tutorial onneural networks,” CoRR, vol. abs/1710.02913, 2017.

[7] S. Marsland, Machine Learning: An Algorithmic Perspective, SecondEdition. Chapman and Hall/CRC, 2014.

[8] A. Burkov, The Hundred-Page Machine Learning Book. Andriy Burkov,2019.

[9] S. Shalev-Shwartz and S. Ben-David, Understanding Machine Learning:From Theory to Algorithms. Cambridge University Press, 2014.

[10] S. O. Haykin, Neural Networks and Learning Machines. Prentice Hall,2008.

[11] K. Crammer and Y. Singer, “On the algorithmic implementation ofmulticlass kernel-based vector machines,” Journal of Machine LearningResearch, vol. 2, p. 265–292, 2001.

[12] L. Hyafil and R. L. Rivest, “Constructing optimal binary decision treesis NP-complete,” Information Processing Letters, vol. 5, no. 1, p. 15–17,1976.

[13] Y. Freund and R. Schapire, “A decision-theoretic generalization of on-line learning and an application to boosting,” European Conference onComputational Learning Theory (EuroCOLT), p. 23–37, 1995.

[14] C. Gianfagna, M. Swaminathan, P. M. Raj, R. Tummala and G. An-tonini, “Enabling antenna design with nano-magnetic materials usingmachine learning,” 2015 IEEE Nanotechnology Materials and DevicesConference (NMDC), pp. 1–5, Anchorage, AK, 2015.

[15] L. Tenuti, G. Oliveri, D. Bresciani and A. Massa, “Advanced learning-based approaches for reflectarrays design,” 2017 11th European Confer-ence on Antennas and Propagation (EUCAP), pp. 84–87, Paris, 2017.

[16] X. Wang, P. Wang and X. Wang, “Adaptive Sparse Array Reconfigura-tion based on Machine Learning Algorithms,” 2018 IEEE InternationalConference on Acoustics, Speech and Signal Processing (ICASSP),pp. 1159–1163, Calgary, AB, 2018.

[17] J. Joung, “Machine Learning-Based Antenna Selection in Wireless Com-munications,” IEEE Communications Letters, vol. 20, no. 11, pp. 2241–2244, Nov. 2016.

[18] Z. M. Fadlullah and M. M. Fouda and N. Kato and X. Shen and Y.Nozaki, “An early warning system against malicious activities for smartgrid communications,” IEEE Network, vol. 25, no. 5, pp. 50–55, 2011.

[19] B. A. Sonkamble and D. D. Doye, “Use of Support Vector Machinesthrough Linear-Polynomial (LP) Kernel for Speech Recognition,” Inter-national Conference on Advances in Mobile Network, Communicationand Its Applications, pp. 46–49, 2012.

[20] C. Perera and A. Zaslavsky and P. Christen and D. Georgakopoulos,“Context Aware Computing for The Internet of Things: A Survey,” IEEECommunications Surveys Tutorials, vol. 16, no. 1, pp. 414–454, 2014.

[21] S. S. W. Samek and T. Wiegand, “The convergence of machine learningand communications,” CoRR, vol. abs/1708.08299, 2017.

[22] J. Huang and J. Antonio Encinar, Reflectarray Antennas. Wiley-IEEEPress, 2007.

[23] D. R. Prado, M. Arrebola, M. R. Pino, R. Florencio, R. R. Boix, J. A.Encinar and F. Las-Heras, “Efficient Crosspolar Optimization of Shaped-Beam Dual-Polarized Reflectarrays Using Full- Wave Analysis for theAntenna Element Characterization,” IEEE Transactions on Antennas andPropagation, vol. 65, no. 2, pp. 623–635, 2017.

[24] P. Robustillo, J. Zapata, J. A. Encinar and J. Rubio, “ANN Charac-terization of Multi-Layer Reflectarray Elements for Contoured-BeamSpace Antennas in the Ku-Band,” IEEE Transactions on Antennas andPropagation, vol. 60, no. 7, pp. 3205–3214, 2012.

[25] V. Richard R., Loison, R. Gillard, H. Legay and M. Romier, “Lossanalysis of a reflectarray cell using ANNs with accurate magnitude pre-diction,” 2017 11th European Conference on Antennas and Propagation(EUCAP), pp. 2396–2399, 2017.

[26] D. R. Prado, J. A. Lopez-Fernandez, M. Arrebola and G. Goussetis,“Efficient Shaped-Beam Reflectarray Design Using Machine LearningTechniques,” 2018 15th European Radar Conference (EuRAD), pp. 525–528, Madrid, 2018.

[27] R. Florencio, R. R. Boix and J. A. Encinar, “Enhanced MoM Analysisof the Scattering by Periodic Strip Gratings in Multilayered Substrates,”IEEE Transactions on Antennas and Propagation, vol. 61, no. 10,pp. 5088–5099, 2013.

[28] N. T. Tokan and F. Gune, “Support Vector Characterization of theMicrostrip Antennas Based on Measurements,” Progress In Electromag-netics Research B, vol. 5, pp. 49–61, 2008.

[29] C. Gianfagna, H. Yu, M. Swaminathan, R. Pulugurtha, R. Tum-mala, and G. Antonini, “Machine-Learning Approach for Designof Nanomagnetic-Based Antennas,” Journal of Electronic Materials,vol. 46, no. 8, pp. 4963–4975, 2017.

[30] J. Tak, A. Kantemur, Y. Sharma and H. Xin, “A 3-D-PrintedW-BandSlotted Waveguide Array Antenna Optimized Using Machine Learning,”IEEE Antennas and Wireless Propagation Letters, vol. 17, no. 11,pp. 2008–2012, Nov. 2018.

[31] A. Hoorfar, “Evolutionary Programming in Electromagnetic Optimiza-tion: A Review,” IEEE Transactions on Antennas and Propagation,vol. 55, no. 3, pp. 523–537, March 2007.

[32] J. M. Johnson and Y. Rahmat-Samii, “Genetic algorithms and methodof moments (GA/MOM) for the design of integrated antennas,” IEEETransactions on Antennas and Propagation, vol. 47, no. 10, pp. 1606–1614, Oct. 1999.

[33] M. D. Gregory, Z. Bayraktar and D. H. Werner, “Fast Optimizationof Electromagnetic Design Problems Using the Covariance MatrixAdaptation Evolutionary Strategy,” IEEE Transactions on Antennas andPropagation, vol. 59, no. 4, pp. 1275–1285, April 2011.

[34] J. Zhang, Z. Zhan, Y. Lin, N. Chen, Y. Gong, J. Zhong, H. S. H. Chung,Y. Li and Y. Shi, “Evolutionary Computation Meets Machine Learning:A Survey,” IEEE Computational Intelligence Magazine, vol. 6, no. 4,pp. 68–75, Nov. 2011.

[35] S. K. Jain, “Bandwidth Enhancement Of Patch Antennas Using Neu-ral Network Dependent Modified Optimizer,” International Journal ofMicrowave and Wireless Technologies, vol. 8, no. 7, pp. 1111–1119,2016.

[36] X. H. Chen, X. X. Guo, J. M. Pei and W. Y. Man, “A hybrid algorithm ofdifferential evolution and machine learning for electromagnetic structureoptimization,” 2017 32nd Youth Academic Annual Conference of ChineseAssociation of Automation (YAC), pp. 755–759, Hefei, 2017.

[37] M. D. Gregory and Z. Bayraktar and D. H. Werner, “Fast Optimizationof Electromagnetic Design Problems Using the Covariance MatrixAdaptation Evolutionary Strategy,” IEEE Transactions on Antennas andPropagation, vol. 59, no. 4, pp. 1275–1285, 2011.

[38] Z. Bayraktar and M. Komurcu and J. A. Bossard and D. H. Werner,“The Wind Driven Optimization Technique and its Application inElectromagnetics,” IEEE Transactions on Antennas and Propagation,vol. 61, no. 5, pp. 2745–2757, 2013.

[39] C. R. Silva and S. R. Martins, “An Adaptive Evolutionary Algorithmfor UWB Microstrip Antennas Optimization Using a Machine LearningTechnique.,” Microw. Opt. Technol. Lett., vol. 55, no. 8, pp. 1864–1868,2013.

[40] B. Liu, H. Aliakbarian, Z. Ma, G. A. E. Vandenbosch, G. Gielen and P.Excell, “An Efficient Method for Antenna Design Optimization Basedon Evolutionary Computation and Machine Learning Techniques,” IEEETransactions on Antennas and Propagation, vol. 62, no. 1, pp. 7–18, Jan.2014.

[41] Saifullah and B. Ahmed, “Machine learning for isotropic antennadesign,” 2018 22nd International Microwave and Radar Conference(MIKON), pp. 683–687, Poznan, 2018.

[42] N. T. Tokan and F. Gunes, “Support vector design of the microstripantenna,” 2008 IEEE 16th Signal Processing, Communication andApplications Conference, pp. 1–4, Aydin, 2008.