Hybrid-Surrogate-Model-Based Efficient Global Optimization for High-Dimensional Antenna Design

Chen, Ling‐Lu; Liao, Cheng; Lin, Wenjun; Chang, Lei; Zhong, and Xuan-Ming

doi:10.2528/pier11121203

Cited by 13 publications

(6 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, numerical experiments show that EGO does not perform as well as other methods on some low-dimensional benchmark problems with steep and narrow global minimum basins (Regis and Shoemaker, 2013) and also on some high-dimensional test problems (e.g. Chen et al 2012).…”

Section: Introductionmentioning

confidence: 97%

Trust regions in Kriging-based optimization with expected improvement

Regis

2015

Engineering Optimization

View full text Add to dashboard Cite

The Kriging-based Efficient Global Optimization (EGO) method works well on many expensive blackbox optimization problems. However, it does not seem to perform well on problems with steep and narrow global minimum basins and on high-dimensional problems. This article develops a new Krigingbased optimization method called TRIKE (Trust Region Implementation in Kriging-based optimization with Expected improvement) that implements a trust-region-like approach where each iterate is obtained by maximizing an Expected Improvement (EI) function within some trust region. This trust region is adjusted depending on the ratio of the actual improvement to the EI. This article also develops the Kriging-based CYCLONE (CYClic Local search in OptimizatioN using Expected improvement) method that uses a cyclic pattern to determine the search regions where the EI is maximized. TRIKE and CYCLONE are compared with EGO on 28 test problems with up to 32 dimensions and on a 36-dimensional groundwater bioremediation application in appendices supplied as an online supplement available at http://dx.. The results show that both algorithms yield substantial improvements over EGO and they are competitive with a radial basis function method.

show abstract

Section: Introductionmentioning

confidence: 97%

Trust regions in Kriging-based optimization with expected improvement

Regis

2015

Engineering Optimization

View full text Add to dashboard Cite

show abstract

“…x c x c (3) where i σ denotes the width of ith hidden neuron and controls the response of the neurons. The training of RBFNN can be briefly described as a process to obtain an approximate functional relationship between inputs and outputs as good as possible in the high-dimensional space [30].…”

Section: Rbf Neural Networkmentioning

confidence: 99%

Fast Multi-Objective Antenna Optimization Based on RBF Neural Network Surrogate Model Optimized by Improved PSO Algorithm

Dong

Wang

2019

Applied Sciences

View full text Add to dashboard Cite

In this paper, a radial basis function neural network (RBFNN) surrogate model optimized by an improved particle swarm optimization (PSO) algorithm is developed to reduce the computation cost of traditional antenna design methods which rely on high-fidelity electromagnetic (EM) simulations. Considering parameters adjustment and update mechanism simultaneously, two modifications are proposed in this improved PSO. First, time-varying learning factors are designed to balance exploration and exploitation ability of particles in the search space. Second, the local best information is added to the updating process of particles except for personal and global best information for better population diversity. The improved PSO is applied to train RBFNN for determining optimal network parameters. As a result, the constructed improved PSO-RBFNN model can be used as a surrogate model for antenna performance prediction with better network generalization capability. By integrating the improved PSO-RBFNN surrogate model with multi-objective evolutionary algorithms (MOEAs), a fast multi-objective antenna optimization framework for multi-parameter antenna structures is then established. Finally, a Pareto-optimal planar miniaturized multiband antenna design is presented, demonstrating that the proposed model provides better prediction performance and considerable computational savings compared to those previously published approaches.

show abstract

“…Compared with traditional EM-driven approaches, surrogate-based optimization techniques construct a mathematical model to predict the antenna performance, thereby greatly reducing the computational cost. Different surrogate models are proposed for antenna designs, such as Kriging [7]- [11], Gaussian Process (GP) [12], [13], and neural networks (NNs) [14]- [17]. The Kriging method used in [7]- [11] is essentially an interpolation method and the model prediction accuracy mostly depends on the initial sample, which may cause the model to either stop prematurely or search too locally [18].…”

Section: Introductionmentioning

confidence: 99%

“…The GP method used in [12] and [13] is still derived from the Kriging model [19] and retains some defects in the Kriging model. Recently, the Neural Network (NN) techniques have also been widely used in antenna designs [14]- [17] to obtain a surrogate model instead of a fine model which has high computational burden. Neural networks learn EM data through the training process and the trained neural networks are then used as fast and accurate surrogate models for complex antenna structure designs.…”

Section: Introductionmentioning

confidence: 99%

Fast Multi-Objective Optimization of Multi-Parameter Antenna Structures Based on Improved BPNN Surrogate Model

2019

View full text Add to dashboard Cite

In this paper, a surrogate model based on a sparsely connected back propagation neural networks (SC-BPNN) is proposed to reduce the large computational cost of conventional multi-objective antenna optimization problems. In this model, the connection parameters and network structure can be adaptively tuned by a hybrid real-binary particle swarm optimization (HPSO) algorithm for better network global optimization capability. Also, a time-varying transfer function is introduced to improve the problem of easily trapping into local optimum and to accelerate network convergence. Further, a fast multi-objective optimization framework based on the proposed SC-BPNN is established for multi-parameter antenna structures. Finally, a Pareto-optimal planar miniaturized multiband antenna design is presented, indicating that the proposed model predicts antenna performance more accurately and saves considerable computational cost compared to those previously published approaches. INDEX TERMS Antenna design, multi-objective optimization, surrogate model, BP neural network, HPSO, transfer function.

show abstract

Hybrid-Surrogate-Model-Based Efficient Global Optimization for High-Dimensional Antenna Design

Cited by 13 publications

References 31 publications

Trust regions in Kriging-based optimization with expected improvement

Trust regions in Kriging-based optimization with expected improvement

Fast Multi-Objective Antenna Optimization Based on RBF Neural Network Surrogate Model Optimized by Improved PSO Algorithm

Fast Multi-Objective Optimization of Multi-Parameter Antenna Structures Based on Improved BPNN Surrogate Model

Contact Info

Product

Resources

About