Search

Hyperspectral band selection algorithms are crucial for processing high-dimensional data, which enables dimensionality reduction, improves data analysis, and enhances computational efficiency. Among these, attention-based algorithms have gained prominence by ranking bands based on their discriminative capability. However, they require a large number of model parameters, which increases the need for extensive training data. To address this challenge, we propose Band Selection through Discrete Relaxation (BSDR), a novel deep learning-based algorithm. BSDR reduces the number of learnable parameters by focusing solely on the target bands, which are typically far fewer than the original bands, thus resulting in a data-efficient configuration that minimizes training data requirements and reduces training time. The algorithm employs discrete relaxation, transforming the discrete problem of band selection into a continuous optimization task, which enables gradient-based search across the spectral dimension. Through extensive evaluations on three benchmark datasets with varying spectral dimensions and characteristics, BSDR demonstrates superior performance for both regression and classification tasks, achieving up to 25% and 34.6% improvements in overall accuracy, compared to the latest attention-based and traditional algorithms, respectively, while reducing execution time by 96.8% and 97.18%. These findings highlight BSDR's effectiveness in addressing key challenges in hyperspectral band selection. [ABSTRACT FROM AUTHOR]

Hyperspectral band selection algorithms are crucial for processing high-dimensional data, which enables dimensionality reduction, improves data analysis, and enhances computational efficiency. Among these, attention-based algorithms have gained prominence by ranking bands based on their discriminative capability. However, they require a large number of model parameters, which increases the need for extensive training data. To address this challenge, we propose Band Selection through Discrete Relaxation (BSDR), a novel deep learning-based algorithm. BSDR reduces the number of learnable parameters by focusing solely on the target bands, which are typically far fewer than the original bands, thus resulting in a data-efficient configuration that minimizes training data requirements and reduces training time. The algorithm employs discrete relaxation, transforming the discrete problem of band selection into a continuous optimization task, which enables gradient-based search across the spectral dimension. Through extensive evaluations on three benchmark datasets with varying spectral dimensions and characteristics, BSDR demonstrates superior performance for both regression and classification tasks, achieving up to 25% and 34.6% improvements in overall accuracy, compared to the latest attention-based and traditional algorithms, respectively, while reducing execution time by 96.8% and 97.18%. These findings highlight BSDR’s effectiveness in addressing key challenges in hyperspectral band selection.

The role of numerical optimization has been continuously growing in the design of high-frequency structures, including microwave and antenna components. At the same time, accurate evaluation of electrical characteristics necessitates full-wave electromagnetic (EM) analysis, which is CPU intensive, especially for complex systems. As rigorous optimization routines involve repetitive EM simulations, the associated cost may be significant. In the design practice, the most widely used EM-driven procedures are by far local (e.g., gradient-based) ones. While typically incurring acceptable expenses that range from dozens to a few hundreds of objective function evaluations, they are prone to failure whenever a decent initial design is not available. Representative scenarios include simulation-based size reduction of compact devices or re-design of structures for operating/material parameters being distant from those at the available design. A standard mitigation approach is the involvement of global search methods, which entails significantly higher computational costs. This paper reviews the recent methodologies introduced to improve the reliability of local parameter tuning algorithms without degrading their computational efficiency. We discuss frequency-based regularization, adaptively adjusted design specification approach, as well as accelerated feature-based optimization. All of these techniques incorporate mechanisms that improve the performance of the search process under challenging scenarios, primarily poor initial conditions. The outline of the mentioned methods is accompanied by illustrative examples including passive microwave circuits and microstrip antennas. Benchmarking against conventional local search is provided as well. Furthermore, the paper discusses the advantages and disadvantages of the reviewed frameworks as well as speculates about future research directions.

Parameter adjustment through numerical optimization has become a commonplace of contemporary microwave engineering. Although circuit theory methods are ubiquitous in the development of microwave components, the initial designs obtained with such tools have to be further tuned to improve the system performance. This is particularly pertinent to miniaturized structures, where the cross-coupling effects cannot be adequately accounted for using equivalent networks. For the sake of reliability, design closure is normally performed using full-wave electromagnetic (EM) simulation models, which entails considerable computational expenses, often impractically excessive. Available mitigation techniques include acceleration of the conventional (e.g., gradient-based) routines using adjoint sensitivities or sparse sensitivity updates, surrogate-assisted and machine learning algorithms, the latter often combined with nature-inspired procedures. Another alternative is the employment of variable-fidelity simulations (e.g., space mapping, co-kriging), which is most often limited to two levels of accuracy (coarse/fine). This work discusses an EM model management approach coupled with trust-region gradient-based routine, which exploits problem-specific knowledge for continuous (multi-level) modification of the discretization density of the microwave structure at hand in the course of the optimization run. The optimization process is launched at the lowest discretization level, thereby allowing for low-cost exploitation of the knowledge about the device under study. Subsequently, based on the convergence indicators, the model fidelity is gradually increased to ensure reliability. The simulation fidelity selection is governed by the algorithm convergence indicators. Computational speedup (i.e., reduction in the number of EM simulations required by the optimization process to converge) is achieved by maintaining low resolution in the initial stages of the optimization run, whereas design quality is secured by eventually switching to the high-fidelity model when close to concluding the process. Numerical verification is carried out using two microstrip circuits, a dual-band power divider and a dual-band branch-line coupler, with the average savings of almost sixty percent when compared to single-fidelity optimization.

Simulation-driven miniaturization of antenna components is a challenging task mainly due to the presence of expensive constraints, evaluation of which involves full-wave electromagnetic (EM) analysis. The recommended approach is implicit constraint handling using penalty functions, which, however, requires a meticulous selection of penalty coefficients, instrumental in ensuring optimization process reliability. This paper proposes a novel size reduction technique with explicit handling of design constraints. Our approach employs trust-region gradient-based procedure as an underlying optimization engine, and adjusts the search radius based on assessing the quality of representing constraints by auxiliary linear expansion models (versus their exact evaluation through EM analysis), rather than based on the quality of objective rendition. This unconventional utilization of trust region framework leads to a procedure that is demonstrably superior over implicit methods, as indicated by comprehensive numerical studies involving three broadband antennas, and benchmarking against state-of-the-art techniques. The two most attractive features of our methodology are simplicity, and no need to tune the algorithm for a specific problem at hand.

Parametric optimization is a mandatory step in the design of contemporary antenna structures. Conceptual development can only provide rough initial designs that have to be further tuned, often extensively. Given the topological complexity of modern antennas, the design closure necessarily involves full-wave electromagnetic (EM) simulations and-in many cases-global search procedures. Both factors make antenna optimization a computationally expensive endeavor: population-based metaheuristics, routinely used in this context, entail significant computational overhead. This letter proposes a novel approach that interleaves trust-region gradient search with iterative parameter space exploration by means of local kriging surrogate models. Dictated by efficiency, the latter are rendered in low-dimensional subspaces spanned by the principal components of the antenna response Jacobian matrix, extracted to identify the directions of the maximum (frequency-averaged) response variability. The aforementioned combination of techniques enables quasi-global search at the cost comparable to local optimization. These features are demonstrated using two antenna examples as well as benchmarking against multiple-start local tuning.

Electromagnetic (EM)-driven optimization is an important part of microwave design, especially for miniaturized components where the cross-coupling effects in tightly arranged layouts make traditional (e.g., equivalent network) representations grossly inaccurate. Efficient parameter tuning requires reasonably good initial designs, which are difficult to be rendered for newly developed structures or when re-design for different operating conditions or material parameters is required. If global search is needed, due to either the aforementioned issues or multi-modality of the objective function, the computational cost of the EM-driven design increases tremendously. This paper introduces a frequency-related regularization as a way of improving the efficacy of simulation-based design processes. Regularization is realized by enhancing the conventional (e.g., minimax) objective function using a dedicated penalty term that fosters the alignment of the circuit characteristics (e.g., the operating frequency or bandwidth) with the target values specified by the design requirement. This leads to smoothening of the objective function landscape, improves reliability of the optimization process, and reduces its computational cost as compared to the standard formulation. An added benefit is the increased immunity to poor initial designs and multi-modality issues. In particular, regularization can make local search routines sufficient in situations where global optimization would normally be necessary. The presented approach is validated using two miniaturized circuits, a rat-race and a branch line coupler. The numerical results demonstrate its superiority over conventional design problem formulations in terms of reliability of the optimization process.

Meticulous tuning of geometry parameters turns pivotal in improving performance of antenna systems. It is more and more often realized using formal optimization methods, which is demonstrably the most efficient way of handling multiple design variables, objectives, and constraints. Although in some cases a need for launching global search arises, a typical design scenario only requires local optimization, especially when a decent initial design can be rendered using engineering experience or parametric studies. At the same time, antenna optimization is typically conducted using full-wave electromagnetic (EM) simulations, which entails considerable computational expenses. In this paper, we introduce a novel procedure for expedited antenna tuning. Its fundamental mechanism is to restrict the antenna response sensitivity updates to the selected dominant directions within the parameter space, determined based on the problem-specific knowledge, in particular, the estimated changes of antenna characteristics when moving across one-dimensional affine subspaces spanned by these directions. Thus, the said dominant directions affect the most the responses of the antenna structure under design as assessed using the introduced metrics. The decision making process concerning the number of directions to be used relies on quantification of the aggregated system response variability metrics. The proposed approach is demonstrated by means of several antenna structures and benchmarked against conventional trust-region algorithm, but also its accelerated versions. The results indicate considerable (up to over 60%) speedup over the reference procedure without noticeable quality degradation. [ABSTRACT FROM AUTHOR]

Supersonic vehicles are an important type of potential transports. Analysis of these vehicles requires the use of accurate models, which are also computationally expensive, to capture the highly nonlinear physics. This paper presents results of numerical investigations of using physics-based surrogate models to design supersonic airfoil shapes. Variable-fidelity models are generated using inviscid computational fluid dynamics simulations and analytical models. By using response correction techniques, in particular, the manifold mapping technique, fast surrogate models are constructed. The effectiveness of the approach is investigated using lift-constrained drag minimization problems of supersonic airfoil shapes. Compared with direct optimization, the results show that an order of magnitude speed up can be obtained. Furthermore, we investigate the effectiveness of the variable-fidelity technique in terms of speed and design quality using several combinations of medium-fidelity and low-fidelity models. [ABSTRACT FROM AUTHOR]

Numerical optimization has become more important than ever in the design of microwave components and systems, primarily as a consequence of increasing performance demands and growing complexity of the circuits. As the parameter tuning is more and more often executed using full-wave electromagnetic (EM) models, the CPU cost of the overall process tends to be excessive even for local optimization. Some ways of alleviating these issues exist, yet, they are limited either by their accessibility or applicability range. This work presents a novel algorithmic approach to accelerated gradient-based parameter tuning of microwave components with numerical derivatives. In our methodology, computational savings are achieved by exploiting the problem-specific knowledge, specifically, by restricting the gradient updates to an orthonormal basis of essential directions corresponding to the maximum variability of the circuit responses within the frequency bands of interest. The said directions are selected through an automated decision-making process involving the analysis of the circuit response variability. Our approach is demonstrated using two multi-parameter microwave devices. Comprehensive comparison with the benchmark methods, including the standard trust-region algorithm and the three accelerated versions, indicate savings of up to fifty percent associated only with minor reduction of the design quality. • Fast algorithm for microwave design with principal directions is developed. • Sensitivity updates are restricted to directions of maximum response changes. • Principal directions are selected through an automated decision-making process. • Proposed approach is demonstrated to yield significant computational savings. • Efficiency improvements do not compromise optimization process reliability. [ABSTRACT FROM AUTHOR]

In this study, a distributed traffic‐aware routing scheme is proposed for a wireless sensor network with multiple sinks. In multi‐sink networks, traffic moving towards one sink may congest that moving towards others. The algorithm proposed here considers the traffic of surrounding neighbours before jumping to any sink. This is accomplished by building for each single sink, a gradient field and using gradient‐based search for routing. The gradient index of one node contains two parts. (i) The first represents the distance‐cost to travel, based on routing over lossy link) proposed by an IETF working group. (ii) The second contains the implicit traffic delay carried by the current holding node. Gradient field is built in distributed way and proven to free from loops. In a simulation, the authors show that this method of building gradient field routes balances the shortest path and possible congestion made by sinks, especially in heavy traffic networks, thereby reducing the overall end‐to‐end delay. [ABSTRACT FROM AUTHOR]

We discuss a simple and yet robust algorithm for electromagnetic (EM) simulation-driven design optimization of antennas. Our procedure exploits adjoint sensitivities to obtain gradient information without extra EM simulations as well as adaptive algorithm that adjusts the search radius for optimizing the linear model of the antenna. The proposed technique is capable of yielding an optimized antenna design at a much lower computational cost than a benchmark algorithm, Matlab's fminimax, also using adjoint sensitivities. Our approach is demonstrated using two antenna design cases: a planar inverted-F antenna and a wideband hybrid antenna. The design time savings are between 60 and 75% with respect to the benchmark method. © 2012 Wiley Periodicals, Inc. Microwave Opt Technol Lett 54:2594-2597, 2012; this online at wileyonlinelibrary.com. DOI 10.1002/mop.27112 [ABSTRACT FROM AUTHOR]

Abstract: This paper examines the use of a so-called “generalised Hammerstein–Wiener” model structure that is formed as the concatenation of an arbitrary number of Hammerstein systems. The latter are taken here to be memoryless non-linearities followed by linear time invariant dynamics. Hammerstein, Wiener, Hammerstein–Wiener and Wiener–Hammerstein models are all special cases of this structure. The parameter estimation of this model is investigated using a standard prediction error criterion coupled with a robust gradient based search algorithm. This approach is profiled using a Wiener–Hammerstein Benchmark example, which illustrates it to be effective and, via Monte-Carlo simulation, relatively robust against capture in local minima. [Copyright &y& Elsevier]

The paper proposes a combination of a gradient-based path planning method and an artificial fish school based supplement for generation of a global optimal path. The optimization criterion has to guarantee „minimum path length and no collisions with obstacles". The main contribution of the paper consists in overcoming the local minima problem by self-organization of an artificial fish school in the current local minimum and its vicinity. The good performance of the proposed modified algorithm is confirmed in MATLAB simulations. [ABSTRACT FROM AUTHOR]

The paper proposes a path-planning approach that generates a global optimal path according to a criterion, which has to guarantee „minimum path length and no collisions with obstacles". The optimization procedures reported in the literature use gradient descent techniques and suffer the local minima problem. The main contribution of the paper consists in overcoming this drawback by neural network based filling the local minimum and its vicinity. The good performance of the proposed modified algorithm is confirmed in MATLAB simulations. [ABSTRACT FROM AUTHOR]

The steepest descent search algorithm is modified in conjunction with chaos to solve the optimization problem of an unstructured search space. The problem is that given only the gradient information of the quality function at the present configuration, X( t), we must find the value of a configuration vector that minimizes the quality function. The proposed algorithm starts basically from the steepest descent search technique but at the prescribed points, i.e., local minimum points, the chaotic jump is performed by the dynamics of a chaotic neuron. Chaotic motions are mainly caused because the Gaussian function has a hysteresis as a refractoriness. An adaptation mechanism to adjust the size of the chaotic jump is also given. In order to enhance the probability of finding the global minimum, a parallel search strategy is developed. The validity of the proposed method is verified in simulation examples of the function minimization problem and the motion planning problem of a mobile robot. [ABSTRACT FROM AUTHOR]

Full‐wave electromagnetic (EM) simulation models are ubiquitous in carrying out design closure of antenna structures. Yet, EM‐based design is expensive due to a large number of analyses necessary to yield an optimized design. Computational savings can be achieved using, for example, adjoint sensitivities, surrogate‐assisted procedures, design space dimensionality reduction, or similar sophisticated means. In this article, a simple modification of a rudimentary trust‐region‐embedded gradient search with numerical derivatives is proposed for reduced‐cost optimization of input characteristics of wideband antennas. The approach exploits information and history of relative changes of the design (as compared with the trust region size) during algorithm iterations to control the updates of components of the antenna response Jacobian, specifically, to execute them only if necessary. It is demonstrated that the proposed framework may lead to over 50% savings over the reference algorithm with only minor degradation of the design quality, specifically, up to 0.3 dB (or <3%). Numerical results are supported by experimental validation of the optimized antenna designs. The presented algorithm can be utilized as a stand‐alone optimization routine or as a building block of surrogate‐assisted procedures. [ABSTRACT FROM AUTHOR]