Optimization Strategy Research Articles

Bayesian view on the training of invertible residual networks for solving linear inverse problems *

Learning-based methods for inverse problems, adapting to the data’s inherent structure, have become ubiquitous in the last decade. Besides empirical investigations of their often remarkable performance, an increasing number of works address the issue of theoretical guarantees. Recently, Arndt et al (2023 Inverse Problems 39 125018) exploited invertible residual networks (iResNets) to learn provably convergent regularizations given reasonable assumptions. They enforced these guarantees by approximating the linear forward operator with an iResNet. Supervised training on relevant samples introduces data dependency into the approach. An open question in this context is to which extent the data’s inherent structure influences the training outcome, i.e. the learned reconstruction scheme. Here, we address this delicate interplay of training design and data dependency from a Bayesian perspective and shed light on opportunities and limitations. We resolve these limitations by analyzing reconstruction-based training of the inverses of iResNets, where we show that this optimization strategy introduces a level of data-dependency that cannot be achieved by approximation training. We further provide and discuss a series of numerical experiments underpinning and extending the theoretical findings.

Recent Advances in Light-Driven Semiconductor-Based Micro/Nanomotors: Optimization Strategies and Emerging Applications.

Micro/nanomotors represent a burgeoning field of research featuring small devices capable of autonomous movement in liquid environments through catalytic reactions and/or external stimuli. This review delves into recent advancements in light-driven semiconductor-based micro/nanomotors (LDSM), focusing on optimized syntheses, enhanced motion mechanisms, and emerging applications in the environmental and biomedical domains. The survey commences with a theoretical introduction to micromotors and their propulsion mechanisms, followed by an exploration of commonly studied LDSM, emphasizing their advantages. Critical properties affecting propulsion, such as surface features, morphology, and size, are presented alongside discussions on external conditions related to light sources and intensity, which are crucial for optimizing the propulsion speed. Each property is accompanied by a theoretical background and conclusions drawn up to 2018. The review further investigates recent adaptations of LDSM, uncovering underlying mechanisms and associated benefits. A brief discussion is included on potential synergistic effects between different external conditions, aiming to enhance efficiency-a relatively underexplored topic. In conclusion, the review outlines emerging applications in biomedicine and environmental monitoring/remediation resulting from recent LDSM research, highlighting the growing significance of this field. The comprehensive exploration of LDSM advancements provides valuable insights for researchers and practitioners seeking to leverage these innovative micro/nanomotors in diverse applications.

Accurate Identification for CW Direct Signal in Underwater Acoustic Ranging

The underwater channel is bilateral, heterogeneous, uncertain, and exhibits multipath transmission, sound line curvature, etc. These properties complicate the structure of the received pulse, causing great challenges in direct signal identification for ranging purposes and impacts on back-end data processing, even accurate acoustic positioning. Machine learning (ML) combined with underwater acoustics has emerged as a prominent area of research in recent years. From a statistical perspective, ML can be viewed as an optimization strategy. Nevertheless, the existing ML-based direct-signal discrimination approaches rely on independent assessment, utilizing a single sensor (beacon or buoy), which is still insufficient for adapting to the complex underwater environment. Thus, discrimination accuracy decreases. To address the above issues, an accurate CW direct signal detection approach is performed using the decision tree algorithm, which belongs to ML. Initially, the pulse parameter characteristics in the underwater multipath channel are investigated and the parameter models are built. Then, based on multi-sensor localization performance feedback, fusion characteristics for diverse pulse are created. Next, the pulse parameter characteristics are preprocessed to mitigate the impact of varying magnitudes and units of magnitude on data processing. Then, the decision tree is built to obtain the desired output results and realize accurate recognition of the ranging direct signals. Finally, the feasibility and reliability of this paper’s method are verified by computer simulation and field testing.

Application of response surface methodology and neural networks in pyramid solar still for seawater desalination: An optimization and prediction strategy

<p>Potable water is essential in various aspects of daily life. Converting brackish water to drinkable water using traditional methods is costly and has environmental impacts. Solar energy is preferred over fossil fuels and other energy sources due to its cost and environmental benefits. Solar stills are increasingly popular due to the growing need for drinkable water, but their performance requires improvement. The study examines the pyramid solar still (PSS) productivity using different operating parameters, such as solar intensity (350-950 W/m2), water inlet temperature (30-50°C), water depth (4-8 cm), and insulation thickness (0.05-0.15m). The response surface technique (RSM) was used to examine the performance of the still under different operating conditions. The RSM approach optimized the process for maximum output and reduced testing time and effort. To verify the optimum response derived from RSM, the artificial neural networks (ANN) model was implemented. The comparative studies of the ANN and RSM models demonstrates a strong correlation with the outcomes achieved in maximizing the productivity of PSS. It found that the ideal values for solar intensity, water intake temperature, water depth, and insulation thickness are 950 W/m2, 50°C, 4 cm, and 0.15 m, respectively. These parameters contribute to the production of 2.585 kg/m2 of distillate.</p>

Optimising building envelope and insulation materials for energy efficiency and life-cycle cost in Cypriot residences

ABSTRACT Residential buildings consume approximately 36% of total electricity consumption in Cyprus and this is expected to continue increasing in the coming years. This study explores the potential for energy savings in residential buildings in Cyprus through optimising building envelopes and insulation materials. Using building information modelling and energy modelling software, the study analyses 120 combinations of building envelopes and eight insulation materials, identifying an optimal combination of wall, roof, and flooring materials that could result in 32% energy load reductions and 27% lower life-cycle costs compared to the median scenario. High-density rock wool insulation on roofs and double-glazed windows were also found to offer both monetary and energy performance benefits. A sensitivity analysis revealed that the current high energy prices could provide greater economic incentives to adopt these optimisation strategies. Optimising building materials in Cyprus could play a critical role in climate strategy while also offering significant economic benefits.

Real-time semantic segmentation via mutual optimization of spatial details and semantic information

Semantic segmentation is a pivotal task in the field of computer vision, encompassing diverse applications and undergoing continuous development. Despite the growing dominance of deep learning methods in this field, many existing network models suffer from trade-offs between accuracy and computational cost, or between speed and accuracy. In essence, semantic segmentation aims to extract semantic information from deep features and optimize them before upsampling output. However, shallow features tend to contain more detailed information but also more noise, while deep features have strong semantic information but lose some spatial information. To address this issue, we propose a novel mutual optimization strategy based on shallow spatial information and deep semantic information, and construct a details and semantic mutual optimization network (DSMONet). This effectively reduces the noise in the shallow features and guides the deep features to reconstruct the lost spatial information, avoiding cumbersome side auxiliary or complex decoders. The Mutual Optimization Module (MOM) includes Semantic Adjustment Details Module (SADM) and Detail Guided Semantic Module (DGSM), which enables mutual optimization of shallow spatial information and deep semantic information. Comparative evaluations against other methods demonstrate that DSMONet achieves a favorable balance between accuracy and speed. On the Cityscapes dataset, DSMONet achieves performances of 79.3% mean of class-wise intersection-over-union (mIoU)/44.6 frames per second (FPS) and 78.0% mIoU/102 FPS. The code is available at https://github.com/m828/DSMONet.

Modelling of patient journey in chronic spontaneous urticaria: Increasing awareness and education by shorten patients' disease journey in Germany.

Chronic spontaneous urticaria (CSU) is both physically and emotionally stressful, and guideline recommendations are often not optimally implemented in clinical practice. The objective of this study was to provide an overview on the patient journey in CSU and to develop a mathematical model based on solid data. The journey of CSU patients in Germany was traced through literature review and expert meetings that included medical experts, pharmacists and representatives of patient organizations. The current situation's main challenges in the patient journey (education, collaboration and disease management) were discussed in depth. Then, a probabilistic model was developed in a co-creation approach to simulate the impact of three potential improvement strategies: (1) patient education campaign, (2) medical professional education programme and (3) implementation of a disease management programme (DMP). Chronic spontaneous urticaria patients are severely burdened by delays in diagnosis and optimal medical care. Our simulation indicates that in Germany, it takes on average of 3.8 years for patients to achieve disease control in Germany. Modelling all three optimization strategies resulted in a reduction to 2.5 years until CSU symptom control. On a population level, the proportion of CSU patients with disease control increased from 44.2% to 58.1%. In principle, effective CSU medications and a disease-specific guideline are available. However, implementation of recommendations is lagging in practice. The approach of quantitative modelling of the patient journey validates obstacles and shows a clear effect of multiple interventions on the patient journey. The data generated by our simulation can be used to identify strategies for improving patient care. Our approach might helping in understanding and improving the management of patients beyond CSU.

Optimal demand management in the electrical microgrid considering energy market prices uncertainty

The optimal management of electrical energy systems is a major concern for energy providers facing uncertainties. In this work, energy demand management and dispatch in the electricity microgrid are performed using an interval optimization strategy under conditions of electricity price uncertainty. Demand dispatch is implemented through the Demand Response Program (DRP) incentive modeling. As a solution, incentive modeling based on the price of the offer to reduce electricity demand during peak periods is proposed. In order to minimize the effects of electricity price uncertainty, the interval optimization approach is used. The main goal of the proposed approach is to minimize the operational cost, and the epsilon constraint-based fuzzy method was used to solve and handle the problem. Finally, the power microgrid was used in different case studies for the results of the numeric simulations and for the positive effects of the proposed modeling under conditions of uncertainty.

Leveraging machine learning for acid catalyzed steam explosion pretreatment: Towards supporting fermentation by the trade-off between glucose and inhibitors

Pretreatment is essential for enhancing the sugar release from lignocellulose. Acid-catalyzed steam explosion (ACSE), a widely-used pretreatment method, still faces challenges, including inhibitors accumulation, which can be overcome by modeling. Here, artificial neural network models were constructed for sulfuric acid-based ACSE to predict sugars and inhibitors, from 12 variables regarding lignocellulose, acid, and steam explosion. Two expanding applications were demonstrated. Firstly, a constraint-based optimization strategy can provide the optimal ACSE condition for fermentation by considering glucose and the synergistic effect of inhibitors on microbial growth simultaneously. Compared to published works, the strategy led to 94% glucose with 22% inhibitors for corn stover, and 100% glucose with 13% inhibitors for wheat straw. Secondly, transfer learning was employed to model phosphoric acid-based ACSE with high accuracy (MSE 0.004) and low data requirement (∼30% of sulfuric acid-based ACSE). The proposed models and applications offer an effective optimization strategy for ACSE and other pretreatment methods for the following fermentation.

Differential evolution with proration-based mutation strategy and multi-segment mixed parameter setting for numerical optimization

In this paper, a new differential evolution (DE) algorithm is presented for solving global optimization problems by guiding the individuals adaptively to explore different decision regions and setting the control parameters properly for individuals. To consolidate algorithm efficiency and preserve population diversity, we first introduce a proration-based mutation strategy. This strategy dynamically selects promising individuals from population as guiders and proportionately allocates the individuals following them based on their fitness values and unsuccessful updates. Besides, we design a multi-segment mixed parameter setting to provide suitable parameters for individuals, considering feedback information from the population, individual requirements, and parameters' interaction. Additionally, we devise an adaptive population maintaining mechanism to refresh the population by replacing individuals with higher unsuccessful updates and removing those with poor fitness values. Unlike previous DE versions, our new algorithm dynamically and appropriately explores promising areas, refines parameter suitability, and accelerates convergence by updating and removing unpromising individuals. This enhances the algorithm search efficiency and achieves a good balance between exploration and exploitation. Finally, we evaluate the proposed algorithm's performance by comparing it with 17 typical or state-of-the-art algorithms on IEEE CEC2014 and CEC2017 test suites. Experimental results indicate that the proposed algorithm is a more promising optimizer.

Design of a Convolutional Neural Network Accelerator Based on On-Chip Data Reordering

Convolutional neural networks have been widely applied in the field of computer vision. In convolutional neural networks, convolution operations account for more than 90% of the total computational workload. The current mainstream approach to achieving high energy-efficient convolution operations is through dedicated hardware accelerators. Convolution operations involve a significant amount of weights and input feature data. Due to limited on-chip cache space in accelerators, there is a significant amount of off-chip DRAM memory access involved in the computation process. The latency of DRAM access is 20 times higher than that of SRAM, and the energy consumption of DRAM access is 100 times higher than that of multiply–accumulate (MAC) units. It is evident that the “memory wall” and “power wall” issues in neural network computation remain challenging. This paper presents the design of a hardware accelerator for convolutional neural networks. It employs a dataflow optimization strategy based on on-chip data reordering. This strategy improves on-chip data utilization and reduces the frequency of data exchanges between on-chip cache and off-chip DRAM. The experimental results indicate that compared to the accelerator without this strategy, it can reduce data exchange frequency by up to 82.9%.

Perspectives for electromagnetic wave absorption with graphene

Graphene-based electromagnetic wave (EMW) absorbing materials have become a focus of scientific research as the demand for addressing the adverse electromagnetic radiation or pollution. Furthermore, graphene-based EMW absorbers promise richer application scenarios than other carbon materials due to graphene's low density, high electrical conductivity, and thermal stability. Consequently, a summary of relevant studies is required. Simultaneously, the functionalities of graphene materials need to be clarified and the importance of graphene materials needs to be emphasized. This work summarizes the research progress of structural optimization strategies of graphene-based EMW absorbing materials from the micro and macro perspectives. The general EMW absorption mechanism is first briefly explained, and the effects of microstructure, components, and varies of loss mechanisms on the EMW absorbing performance of graphene-based EMW absorbing materials are discussed, in particular, the differences in electromagnetic parameters and absorbing properties caused by anisotropic structures are summarized. Finally, the outlook regarding the research focus and development direction of graphene-based EMW absorbing materials is proposed.

Highly efficient expression of Rasamsonia emersonii lipase in Pichia pastoris: characterization and gastrointestinal simulated digestion in vitro.

Acidic lipases with high catalytic activities under acidic conditions have important application values in the food, feed and pharmaceutical industries. However, the availability of acidic lipases is still the main obstacle to their industrial applications. Although a novel acidic lipase Rasamsonia emersonii (LIPR) was heterologously expressed in Escherichia coli, the expression level was unsatisfactory. To achieve the high-efficiency expression and secretion of LIPR in Pichia pastoris GS115, the combinatorial optimization strategy was adopted including gene codon preference, signal peptide, molecular chaperone co-expression and disruption of vacuolar sorting receptor VPS10. The activity of the combinatorial optimization engineered strain in a shake flask reached 1480 U mL-1, which was 8.13 times greater than the P. pastoris GS115 parental strain. After high-density fermentation in a 5-L bioreactor, the highest enzyme activity reached as high as 11 820 U mL-1. LIPR showed the highest activity at 40 °C and pH 4.0 in the presence of Ca2+ ion. LIPR exhibited strong tolerance to methanol, indicating its potential application in biodiesel biosynthesis. Moreover, the gastrointestinal digestion simulation results demonstrated that LIPR was tolerant to pepsin and trypsin, but its activity was inhibited by sodium taurodeoxycholate. This study provided an effective approach for the high expression of acidic lipase LIPR. LIPR was more appropriate for lipid digestion in the stomach than in intestine according to the gastrointestinal digestion simulation results. © 2024 Society of Chemical Industry.

Optimization of duplex digital PCR for the measurement of SARS-CoV-2 RNA

Quantitative PCR (qPCR) is the gold standard for detecting nucleic acid sequences specific to the target pathogen. For COVID-19 diagnosis, several molecular assays have been developed. In this study, we present an optimization strategy for the measurement of SARS-CoV-2 RNA via multiplex qPCR and digital PCR (dPCR). Compared to qPCR, both droplet and chip-based dPCR, which are known to be more sensitive and accurate, showed a better resilience to suboptimal assay compositions and cycling conditions following the proposed optimizations. In particular, the formation of heterodimers among assays greatly interfered with qPCR results, but only minimally with dPCR results. In dPCR, existing heterodimers lowered the PCR efficiency, producing a dampened fluorescent signal in positive partitions. This can be corrected by adjusting the PCR cycling conditions, after which dPCR shows the capability of measuring the expected copy number. In addition, we present a process to improve the existing RdRp assay by correcting the primer sequences and matching the melting temperature, ultimately producing highly sensitive and robust assays. The results of this study can reduce the cost and time of SARS-CoV-2 diagnosis while increasing accuracy. Furthermore, our results suggest that dPCR is a reliable method for the accurate measurement of nucleic acid targets.

Digital Twin for the SBR Cold Emulsion Copolymerization Process

AbstractDevelopment of accurate and robust dynamic models for the description of emulsion copolymerization processes is fundamental for the implementation of monitoring, advanced control, and optimization strategies. There are several studies on the dynamic modeling of styrene/1,3‐butadiene rubber (SBR) emulsion copolymerization, but most of them focus on hot conditions or only one semi‐batch reactor, as in the case of cold conditions. For this reason, the present study focuses on the dynamic modeling of SBR cold emulsion copolymerization processes considering a train of 15 continuous stirred tank reactors, as in many real industrial sites. The developed dynamic model is implemented by using the digital twin (DT) concept, which involves the online reading of process variables and an adaptive strategy for online tuning of some of the model parameters, being also sensitive to the effect of real‐time changes on the number of reactors in the train, a subject that has been overlooked previously, but which is important at the plant site. The practical application of the DT for monitoring a real industrial process illustrates the robustness and accuracy of the developed tool, making it useful for opportune detection of process anomalies and opening the way for future advanced control strategies.

Model Prediction and Optimization of Cefixime Trihydrate Removal from Simulated Wastewater through Advanced Plasma Technology

Plasma technology has received significant attention as an advanced oxidation technique for the efficient removal of pharmaceuticals from various wastewaters. In this study, cefixime trihydrate (CFX) was used as a typical pharmaceutical, and a stepwise optimization strategy was adopted to determine the experimental parameters that benefiting the generation of free radical. The removal rate of 20 mg/L CFX reached 94.8% after 60 min plasma treatment when the voltage was 12 kV, the discharge gap was 15 mm, and the number of needles was 5. Furthermore, through the prediction and optimization of the response surfaces mythology (RSM) and artificial neural networks (ANN) model, the removal rate of CFX increased to 95.8% when the voltage was 11.9 kV, discharge gap was 15 mm, the number of needles was 5, and treatment time was 60 min. Finally, based upon the change of solution absorbance before and after the UV-Vis scan for plasma treatment and free radical quenching experiments, ·OH, 1O2, and ·O2− mainly attacked the C=C of CFX and played an important role during the plasma treatment. Moreover, the toxicity of CFX solution after ·OH attack was reduced evidently compared with that of the initial CFX solution via further toxicity assessment.

Design and optimization of driving mode control strategies for front-and-rear-axle-independent-drive-type electric vehicle

Design and optimization of driving mode control strategies for front-and-rear-axle-independent-drive-type electric vehicle

Boundary Layer Stagnation Point Flow and Heat Transfer over a Nonlinear Stretching/Shrinking Sheet in Hybrid Carbon Nanotubes: Numerical Analysis and Response Surface Methodology under the Influence of Magnetohydrodynamics

The present study aims to offer new numerical solutions and optimisation strategies for the fluid flow and heat transfer behaviour at a stagnation point through a nonlinear sheet that is expanding or contracting in water-based hybrid nanofluids. Most hybrid nanofluids typically use metallic nanoparticles. However, we deliver a new approach by combining single- and multi-walled carbon nanotubes (SWCNTs-MWCNTs). The flow is presumptively steady, laminar, and surrounded by a constant temperature of the ambient and body walls. By using similarity variables, a model of partial differential equations (PDEs) with the magnetohydrodynamics (MHD) effect on the momentum equation is converted into a model of non-dimensional ordinary differential equations (ODEs). Then, the dimensionless first-order ODEs are solved numerically using the MATLAB R2022b bvp4C program. In order to explore the range of computational solutions and physical quantities, several dimensionless variables are manipulated, including the magnetic parameter, the stretching/shrinking parameter, and the volume fraction parameters of hybrid and mono carbon nanotubes. To enhance the originality and effectiveness of this study for practical applications, we optimise the heat transfer coefficient via the response surface methodology (RSM). We apply a face-centred central composite design (CCF) and perform the CCF using Minitab. All of our findings are presented and illustrated in tabular and graphic form. We have made notable contributions in the disciplines of mathematical analysis and fluid dynamics. From our observations, we find that multiple solutions appear when the magnetic parameter is less than 1. We also detect double solutions in the shrinking region. Furthermore, the increase in the magnetic parameter and SWCNTs-MWCNTs volume fraction parameter increases both the skin friction coefficient and the local Nusselt number. To compare the performance of hybrid nanofluids and mono nanofluids, we note that hybrid nanofluids work better than single nanofluids both in skin friction and heat transfer coefficients.

A self-supervised learning model for graph clustering optimization problems

The graph clustering optimization problems are ubiquitous in both scientific and engineering fields, for example the graph partitioning problem and facility location problem. Since most of these problems are NP-hard, it is challenging to design effective algorithms for them. Recently, deep learning has achieved promising performances in solving combinatorial optimization problems. However, existing supervised learning-based methods are limited by the absence of labeled instances, while reinforcement learning-based methods tend to be hard to train and converge slowly for large instances. In this paper, we propose a self-supervised model named GCOM to address NP-hard graph clustering optimization problems. It adopts a pre-training and fine-tuning framework. In the pre-training process, we utilize graph contrastive learning to mine the latent clustering relations and useful optimization strategies, which can be generalized to the downstream tasks. Additionally, two effective graph encodings are introduced to enhance the expressive power of our model. We evaluate our model on facility location problem (FLP) and balanced graph partitioning (BGP). Extensive experiments conducted on both synthetic and real-world datasets validate superior performance, strong generalization ability and high sample-efficiency of our model.

A study on shaping tourists’ conservational intentions towards cultural heritage in the digital era: exploring the effects of authenticity, cultural experience, and place attachment

ABSTRACT With the rapid development of digital tourism, more and more tourists are embracing this innovative mode of tourism. Digital tourism has attracted a large influx of tourists, which not only promotes the dissemination of local culture, but also promotes tourists’ intention to preserve culture. This study proposes a comprehensive framework authenticity, cultural experience, place attachment, satisfaction, and cultural protection intention, aiming to explore tourists’ cultural protection intention in digital tourism. To validate our framework, we use the Digital Intelligence Niushou as a case study and distributed a questionnaire to tourists, obtaining 408 valid responses. We analyzed the collected data using structural equation modeling (SEM). The results showed that place attachment and the four cultural experiences positively influenced both visitors’ satisfaction and cultural protection intention. Notably, “scientific knowledge” and “personal values” emerged as the biggest influences on cultural protection intention. Among the three types of authenticity, only objective authenticity and constructed authenticity positively influenced tourists’ satisfaction and cultural protection intention. This suggests that faithfully restored historical contexts play a key role in shaping tourists’ experiences and attitudes toward cultural protection in the field of digital tourism. This study provides a solid theoretical foundation for advancing digital tourism and enhancing tourists’ commitment to preserving cultural heritage. Our findings are important for digital tourism platforms to develop more effective optimization strategies, and can help managers to create more attractive, authentic and immersive tourism models to enhance tourist satisfaction and cultivate cultural protection intention.