Metamodel-Assisted Evolution Strategies Applied in Electromagnetic Compatibility Design (original) (raw)
Related papers
A Parallel Surrogate Model Assisted Evolutionary Algorithm for Electromagnetic Design Optimization
IEEE Transactions on Emerging Topics in Computational Intelligence, 2019
Optimization efficiency is a major challenge for electromagnetic (EM) device, circuit and machine design. Although both surrogate model-assisted evolutionary algorithms (SAEAs) and parallel computing are playing important roles in addressing this challenge, there is little research that investigates their integration to benefit from both techniques. In this paper, a new method, called parallel SAEA for electromagnetic design (PSAED), is proposed. A state-of-the-art SAEA framework, surrogate model-aware evolutionary search, is used as the foundation of PSAED. Considering the landscape characteristics of EM design problems, three differential evolution mutation operators are selected and organized in a particular way. A new SAEA framework is then proposed to make use of the selected mutation operators in a parallel computing environment. PSAED is tested by a micromirror and a dielectric resonator antenna as well as four mathematical benchmark problems of various complexity. Comparisons with state-of-the-art methods verify the advantages of PSAED in terms of efficiency and optimization capacity.
Electromagnetic device optimization by hybrid evolution strategy approaches
COMPEL: The International Journal for Computation and Mathematics in Electrical and Electronic Engineering, 2007
Purpose -This paper aims to show on a widely used benchmark problem that chaotic sequences can improve the search ability of evolution strategies (ES). Design/methodology/approach -The Lozi map is used to generate new individuals in the framework of ES algorithms. A quasi-Newton (QN) method is also used within the iterative loop to improve the solution's quality locally. Findings -It is shown that the combined use of chaotic sequences and QN methods can provide high-quality solutions with small standard deviation on the selected benchmark problem. Research limitations/implications -Although the benchmark is considered to be representative of typical electromagnetic problems, different test cases may give less satisfactory results. Practical implications -The proposed approach appears to be an efficient general purpose optimizer for electromagnetic design problems. Originality/value -This paper introduces the use of chaotic sequences in the area of electromagnetic design optimization.
COMPEL: The International Journal for Computation and Mathematics in Electrical and Electronic Engineering, 2009
Purpose -The purpose of this paper is to show, on a widely used benchmark problem, that adaptive mutation factors and attractive/repulsive phases guided by population diversity can improve the search ability of differential evolution (DE) algorithms. Design/methodology/approach -An adaptive mutation factor and attractive/repulsive phases guided by population diversity are used within the framework of DE algorithms. Findings -The paper shows that the combined use of adaptive mutation factors and population diversity in order to guide the attractive/repulsive behavior of DE algorithms can provide high-quality solutions with small standard deviation on the selected benchmark problem. Research limitations/implications -Although the chosen benchmark is considered to be representative of typical electromagnetic problems, different test cases may give less satisfactory results. Practical implications -The proposed approach appears to be an efficient general purpose stochastic optimizer for electromagnetic design problems. Originality/value -This paper introduces the use of population diversity in order to guide the attractive/repulsive behavior of DE algorithms.
COMPEL - The international journal for computation and mathematics in electrical and electronic engineering, 2007
PurposeThe purpose of this paper is to propose a surrogate model‐assisted optimization algorithm which effectively searches for the optimum at the earliest opportunity, avoiding the need for a large initial experimental design, which may be wasteful.Design/methodology/approachThe methodologies of two‐stage and one‐stage selection of points are combined for the first time. After creating a small experimental design, a one‐stage Kriging algorithm is used to search for the optimum for a fixed number of iterations. If it fails to locate the optimum, the points it samples are then used in lieu of a traditional experimental design to initialize a two‐stage algorithm.FindingsThe proposed approach was tested on a mathematical test function. It was found that the optimum could be located, without necessarily constructing an accurate surrogate model first. The algorithm performed well on an electromagnetic design problem, outperforming both a random search and a genetic algorithm, in signific...
2012
Simulation-based optimization has become an important design tool in microwave engineering. Yet, employing electromagnetic (EM) solvers in the design process is a challenging task, primarily due to a high-computational cost of an accurate EM simulation. This paper is focused on efficient EM-driven design optimization techniques that utilize physically-based low-fidelity models, normally based on coarse-discretization EM simulations. The presented methods attempt to exploit as much of the knowledge about the system or device of interest embedded in the low-fidelity model as possible, so as to reduce the computational cost of the design process. Unlike many other surrogate-based approaches, the techniques discussed here are non-parametric ones, i.e., they are not based on analytical formulas. The paper presents several specific methods, including those based on correcting the low-fidelity model response (adaptive response correction and shape-preserving response prediction), as well as on suitable modification of the design specifications. Formulations, application examples and the discussion of advantages and disadvantages of these techniques are also included.
2014
In recent years, various methods from the evolutionary computation (EC) field have been applied to electromagnetic (EM) design problems and have shown promising results. However, due to the high computational cost of the EM simulations, the efficiency of directly using evolutionary algorithms is often very low (e.g., several weeks' optimization time), which limits the application of these methods for many industrial applications. To address this problem, a new method, called Surrogate Model Assisted Differential Evolution for Antenna Synthesis (SADEA), is presented in this paper. The key ideas are: (1) A Gaussian Process (GP) surrogate model is constructed on-line to predict the performances of the candidate designs, saving a lot of computationally expensive EM simulations. (2) A novel surrogate model-aware evolutionary search mechanism is proposed, directing effective global search even when a traditional high-quality surrogate model is not available. Three complex antennas and two mathematical benchmark problems are selected as examples. Compared with the widely used differential evolution and particle swarm optimization, SADEA can obtain comparable results, but achieves a 3 to 7 times speed enhancement for the antenna design optimization.
A hybrid multiobjective differential evolution method for electromagnetic device optimization
COMPEL: The International Journal for Computation and Mathematics in Electrical and Electronic Engineering, 2011
Purpose -The purpose of this paper is to show that the performance of differential evolution (DE) can be substantially improved by a combination of techniques. These enhancements are applicable to both single and multiobjective problems. Their combined use allows the optimization of complex 3D electromagnetic devices. Design/methodology/approach -DE is improved by a combination of techniques which are applied in a cascade way and their single and combined effect is tested on well-known benchmarks and domain-specific applications. Findings -It is shown that the combined use of enhancement techniques provides substantial improvements in the speed of convergence for both single and multiobjective problems.
IET Microwaves, Antennas & Propagation, 2014
In this study, the authors describe a simple and reliable procedure for electromagnetic (EM) simulation-based design of antennas. The authors' approach exploits coarse-discretisation EM simulations (the low-fidelity antenna model) and the adaptive response correction (ARC) technique. ARC is used to enhance the low-fidelity model response so that it can serve as a cheap yet accurate representation of the high-fidelity EM model of the antenna under design. The ARC-based iterative correction-prediction process yields an optimised design at a low CPU cost corresponding to a few high-fidelity simulations of the antenna. The proposed approach is compared with direct optimisation of the high-fidelity model and to surrogate-based optimisation exploiting output space mapping. The results obtained for two antenna design cases indicate by reduction of the overall design cost that ARC allows for better utilisation of the knowledge embedded in the low-fidelity model than the simpler correction method.
2011 IEEE International Symposium on Electromagnetic Compatibility, 2011
We propose a new parametric macromodeling technique for complex electromagnetic systems described by scattering parameters, which are parameterized by multiple design variables such as layout or substrate feature. The proposed technique is based on an efficient and reliable combination of rational identification, a procedure to find scaling and frequency shifting system coefficients, and positive interpolation schemes. Parametric macromodels can be used for efficient and accurate design space exploration and optimization. A design optimization example for a complex electromagnetic system is used to validate the proposed parametric macromodeling technique in a practical design process flow.
Strategies for balancing exploration and exploitation in electromagnetic optimisation
Purpose – Electromagnetic design utilising finite element or similar numerical methods is computationally expensive, thus efficient algorithms reducing the number of objective function calls to locate the optimum are sought. The balance between exploration and exploitation may be achieved using a reinforcement learning approach, as demonstrated previously. However, in practical design problems, in addition to finding the global optimum efficiently, information about the robustness of the solution may also be important. In this paper, the aim is to discuss the suitability of different search algorithms and to present their fitness to solve the optimization problem in conjunction with providing enough information on the robustness of the solution. Design/methodology/approach – Two novel strategies enhanced by the surrogate model based weighted expected improvement approach are discussed. The algorithms are tested using a two-variable test function. The emphasis of these strategies is on accurate approximation of the shape of the objective function to accomplish a robust design. Findings – The two novel strategies aim to pursue the optimal value of weights for exploration and exploitation throughout the iterative process for better prediction of the shape of the objective function. Originality/value – It is argued that the proposed strategies based on adaptively tuning weights perform better in predicting the shape of the objective function. Good accuracy of predicting the shape of the objective function is crucial for achieving a robust design.