Learning hyperparameter predictors for similarity-based multidisciplinary topology optimization.

Topology optimization (TO) plays a significant role in industry by providing engineers with optimal material distributions based exclusively on the information about the design space and loading conditions. Such approaches are especially important for current multidisciplinary design tasks in industry, where the conflicting criteria often lead to very unintuitive solutions. Despite the progress in integrating manufacturing constraints into TO, one of the main factors restricting the use of TO in practice is the users' limited control of the final material distribution. To address this problem, recently, a universal methodology for enforcing similarity to reference structures in various TO methods by applying scaling of elemental energies was proposed. The method, however, requires an expensive hyperparameter sampling, which involves running multiple TO processes to find the design of a given similarity to a reference structure. In this article, we propose a novel end-to-end approach for similarity-based TO, which integrates a machine learning model to predict the hyperparameters of the method, and provide the engineer, at minimal computational cost, with a design satisfying multidisciplinary criteria expressed by the similarity to a reference. The training set for the model is generated based on an academic linear elastic problem, but the model generalizes well to both nonlinear dynamic crash and industrial-scale TO problems. We show the latter by applying the proposed methodology to a real-world multidisciplinary TO problem of a car hood frame, which demonstrates the usefulness of the approach in industrial settings.

Impact of Glucose Measuring Systems and Sample Type on Diagnosis Rates of Diabetes Mellitus.

INTRODUCTION: The use of glucose point-of-care testing (POCT) devices for the diagnosis of diabetes mellitus (DM) is an ongoing controversy. In patient management, glucose concentrations are determined by POCT and core laboratory glucose methods, and the values are commonly compared even though the samples collected are different, namely, capillary whole blood and venous plasma. In individual patients it is difficult to distinguish between factors that can influence the results, such as sample type and measuring procedure. In this study, glucose concentrations obtained using POCT and core laboratory instruments were assessed to duplicate typical scenarios experienced in healthcare. Corresponding diagnosis rates of impaired glucose tolerance (IGT) and DM based on fixed, method-independent cutoffs were compared. METHODS: Glucose concentration was measured by the 2-h oral glucose tolerance test (OGTT) in samples collected from an inpatient cohort and a cohort from the general population. Two POCT methods, namely, a handheld unit-use glucometer and a small bench-top analyzer with batch reagents, and two core laboratory procedures were used to measure glucose concentrations. The sample types were whole blood and plasma samples collected from venous and capillary blood. The glycated hemoglobin level in whole blood was also determined. RESULTS: A total of 231 subjects were included in the study. The 2-h OGTT glucose concentrations in the capillary whole blood samples showed a positive bias of 0.8 mmol/L compared to those obtained using core laboratory plasma glucose methods, leading to increased rates of diabetes diagnosis. This bias decreased to 0.2 mmol/L when venous blood was used in the tests. A change in the method used by the core laboratory introduced a negative bias of 0.5 mmol/L and, consequently, a lower diagnosis rates. DISCUSSION AND CONCLUSION: Venous blood samples measured at the point-of-care are the most suitable sample type for the measurement of the glucose concentration in the 2-h OGTT. The investigated unit-use POCT method is suitable for the diagnosis of IGT and DM when venous blood samples are collected. Importantly, changes in measurement procedures can introduce a bias and affect diagnosis rates, thereby emphasizing the need for further harmonization of glucose methods. A plain language summary is available for this article.

High-power arrays of quantum cascade laser master-oscillator power-amplifiers.

We report on multi-wavelength arrays of master-oscillator power-amplifier quantum cascade lasers operating at wavelengths between 9.2 and 9.8 µm. All elements of the high-performance array feature longitudinal (spectral) as well as transverse single-mode emission at peak powers between 2.7 and 10 W at room temperature. The performance of two arrays that are based on different seed-section designs is thoroughly studied and compared. High output power and excellent beam quality render the arrays highly suitable for stand-off spectroscopy applications.

Evolution by adapting surrogates.

To deal with complex optimization problems plagued with computationally expensive fitness functions, the use of surrogates to replace the original functions within the evolutionary framework is becoming a common practice. However, the appropriate datacentric approximation methodology to use for the construction of surrogate model would depend largely on the nature of the problem of interest, which varies from fitness landscape and state of the evolutionary search, to the characteristics of search algorithm used. This has given rise to the plethora of surrogate-assisted evolutionary frameworks proposed in the literature with ad hoc approximation/surrogate modeling methodologies considered. Since prior knowledge on the suitability of the data centric approximation methodology to use in surrogate-assisted evolutionary optimization is typically unavailable beforehand, this paper presents a novel evolutionary framework with the evolvability learning of surrogates (EvoLS) operating on multiple diverse approximation methodologies in the search. Further, in contrast to the common use of fitness prediction error as a criterion for the selection of surrogates, the concept of evolvability to indicate the productivity or suitability of an approximation methodology that brings about fitness improvement in the evolutionary search is introduced as the basis for adaptation. The backbone of the proposed EvoLS is a statistical learning scheme to determine the evolvability of each approximation methodology while the search progresses online. For each individual solution, the most productive approximation methodology is inferred, that is, the method with highest evolvability measure. Fitness improving surrogates are subsequently constructed for use within a trust-region enabled local search strategy, leading to the self-configuration of a surrogate-assisted memetic algorithm for solving computationally expensive problems. A numerical study of EvoLS on commonly used benchmark problems and a real-world computationally expensive aerodynamic car rear design problem highlights the efficacy of the proposed EvoLS in attaining reliable, high quality, and efficient performance under a limited computational budget.

Widely tunable mid-infrared quantum cascade lasers using sampled grating reflectors.

We demonstrate a three-section, electrically pulsed quantum cascade laser which consists of a Fabry-Pérot section placed between two sampled grating distributed Bragg reflectors. The device is current-tuned between ten single modes spanning a range of 0.46 µm (63 cm(-1)), from 8.32 to 8.78 µm. The peak optical output power exceeds 280 mW for nine of the modes.

Quantum cascade laser master-oscillator power-amplifier with 1.5 W output power at 300 K.

We report quantum cascade laser (QCL) master-oscillator power-amplifiers (MOPAs) at 300 K reaching output power of 1.5 W for tapered devices and 0.9 W for untapered devices. The devices display single-longitudinal-mode emission at λ = 7.26 µm and single-transverse-mode emission at TM(00). The maximum amplification factor is 12 dB for the tapered devices.