Performance Optimization Techniques Research Articles

A novel approach using copper oxide nanocomposites-based biosensor for lactate detection in athletes

Accurate monitoring tools are necessary for lactate, a critical biomarker linked to anaerobic metabolism and muscular performance. Lactate dehydrogenase (LDH) immobilized on MXene (Ti3C2Tx)-supported CuO nanocomposites (MCNs) modified glassy carbon electrode (LDH/MCN/GCE) is the innovative electrochemical biosensor for lactate detection presented in this work. CuO nanostructure integration onto MXene nanosheets, MCN production, and LDH immobilization were all successfully accomplished, according to structural characterization performed using XRD, SEM, XPS, and FTIR. The best lactate oxidation current was shown by LDH/MCN/GCE, according to electrochemical study. With a low detection limit of 0.25 µM and a high sensitivity of 5.92107 µA/mM, the biosensor exhibited a linear response range of 0.1–121 mM. The remarkable selectivity, stability, and sensitivity were attributed to enhanced LDH immobilization made possible by MCNs and MXene. The promise of LDH/MCN/GCE for bioanalytical applications is highlighted by promising results from practical testing in human blood samples, opening the door for improvements in lactate monitoring and performance optimization techniques.

Performance optimisation techniques for Conflict-free Replicated Data Types (CRDT)

The research relevance is determined by the need for distributed data handling optimisation that does not cause conflicts during replication, since the introduction of such types of data leads to an increase in the need for their optimisation and performance. The study aims to develop techniques to improve the efficiency of various types of data with conflict-free replication. To achieve this goal, the methods of analysis and experimentation were used. The study results demonstrate that the use of certain optimised methods of conflict-free data types in replication leads to a significant improvement in efficiency and reduction of resource costs. The results show that caching and precomputation, asynchronous synchronisation, and local processing are the most effective methods for improving the efficiency of Conflict-free Replicated Data Types. It is established that the choice of a specific method for improving the performance of Conflict-free Replicated Data Types should be justified and based on a careful analysis of the requirements and characteristics of a particular distributed system. This takes into account user needs, data criticality, edit frequency, etc. The research includes the development and optimisation of synchronisation algorithms that significantly improve the performance of these types of data in distributed systems. The work includes simple software implementations for detailed analysis and study of methods for optimising the performance of data types in conflict-free replications. The introduction of distributed computing has made it possible to optimise the processes of data replication and synchronisation, which helps to reduce overheads and improve system performance. It is concluded that the study is of practical importance since the use of optimised conflict-free data types with replication in real distributed systems can improve their efficiency and reliability. The practical significance of the study lies in the possibility of applying optimised methods of data types with conflict-free replication in real distributed systems, which will increase their performance and ensure efficient operation under conditions of high load and limited resource potential

Multiscale Bidirectional Temporal Convolutional Network for Sleep Apnea Detection Based on Wearable Photoplethysmography Bracelet.

Sleep apnea syndrome (SAS), which can lead to a range of Cardiopulmonary diseases, is a common chronic sleep disorder. The unobtrusive detection based on wearable devices is helpful for early diagnosis and treatment of SAS. To this end, this paper presents a method based on a one-dimensional multi-scale bidirectional temporal convolutional neural network (1D-MsBiTCNet) and two model performance optimization techniques, i.e., regularized dropout (RD) and logit adjustment (LA). Among them, 1D-MsBiTCNet has outstanding capabilities in both feature extraction and temporal dependence representation. RD and LA play an effective role in solving the overfitting problem of model training and the class imbalance problem of the dataset, respectively. The proposed model was trained and tested on a photoplethysmography (PPG) dataset (including data from 92 subjects) collected from commercial wearable bracelets. On this dataset, our method achieved accuracy, sensitivity and specificity of 82.76%, 71.58%, 86.74% for per-segment detection, and 97.83%, 88.89%, 100.00% for per-recording severe SAS detection. For the precise quantification of apnea-hypopnea index (AHI), our method achieved a mean absolute error of 5.44 between the predicted AHI and the ground truth AHI. The experimental results show that our proposed method has an outstanding performance and can provide a methodological reference for large-scale SAS automatic detection.

Cost-Effective Resource Allocation and Optimization Strategies for Multi-Cloud Environments

Abstract: As organizations increasingly adopt multi-cloud strategies to leverage the benefits of multiple cloud service providers, the need for effective cost optimization becomes paramount. This research paper aims to explore and propose innovative approaches for optimizing costs in multi-cloud environments. The study will investigate resource allocation, workload placement, and performance optimization techniques, taking into account factors such as workload variability, service-level agreements (SLAs), and pricing models across different cloud providers. The objective is to develop cost-effective strategies that balance performance and cost across a heterogeneous cloud ecosystem, ultimately helping organizations make informed decisions in their multi-cloud deployments. This research topic addresses the growing interest in multi-cloud architectures and the associated challenges in managing costs while ensuring high-quality service delivery. It provides a valuable contribution to the field of cloud computing and cost optimization, which is increasingly crucial for businesses and enterprises seeking to maximize the benefits of cloud technology.

범주 불균형 금융 문제의 해결을 위한 부스팅 학습

The class imbalance problem is essential to be solved in order to improve the performance of classification and prediction models. This study proposes an AUCBooost technique that introduces a direct performance optimization technique to improve the performance of boosting algorithms applied to class imbalance problems in the financial sector. In this study, the performance of AUCBoost is verified on class imbalance problems such as corporate bankruptcy, card insolvency, and card fraud. For comparative analysis of performance, logistic, AdaBoost, GBM, and XGBoost are adopted as benchmark models. The results of repeated 10-fold cross-validation are as follows. First, unlike conventional algorithms only focusing on majority class in class imbalance problems, AUCBoost shows a balanced learning behavior that simultaneously considers the specificity of multiple class and the sensitivity of minority class. Second, compared to the benchmarking model, RMA shows that AUCBoost shows significantly improved performance in terms of AUC. Third, in class balanced data generated from data sampling, AUCBooost also shows the significantly improved performance compared to benchmarking models.

Metaheuristic based single and multiobjective optimization of thermoelectric generator

Analytical optimization methods are not applicable for some complex practical problems that lead to explore the numerical/metaheuristic optimization methods. Thermoelectric devices are also complex systems and need metaheuristic methods for their performance optimization. This paper analyses various metaheuristic techniques for performance optimization of thermoelectric generator (TEG). Single and multiobjective optimizations of TEG have been performed to optimize its power output and efficiency. The hot and cold side temperatures, load resistance, length of semiconductor legs, and fill factor of the TEG are used as performance parameters. Single-objective optimization is performed using 13-metaheuristic algorithms and their results are compared in terms of solution optimality and computational time acceptability. It has been observed that the cuckoo search (CS) algorithm is the most suitable in terms of accuracy and speed. CS produced the optimum power output and efficiency of 19.20837 W and 2.4427 % respectively in minimum time. The multiobjective optimization is also performed for simultaneous optimization of the power output and efficiency. Shannon entropy approach is found to be the best technique to determine optimal point on the Pareto front for multiobjective optimization. The optimal values of decision variables, Th Tc, RL, L and FF for multiobjective optimization are found to be 321.195 K, 281.195 K, 0.14250 Ω, 0.5 mm and 0.9 respectively and it produced optimum power output and efficiency of 19.20837 W and 2.25932 % respectively. The statistical and box plot analysis have also been carried out for all the algorithms. The determination of the most suitable algorithm for the faster performance optimization of the TEG, hereby provides the pathways to ascertain the most accurate solution for the power output and efficiency obtainable from the TEG system and hence, delivering the ability to decide the right parameters for the construction of the TEG system with the best performance.

A Tutorial on Agricultural IoT: Fundamental Concepts, Architectures, Routing, and Optimization

This paper presents an in-depth contextualized tutorial on Agricultural IoT (Agri-IoT), covering the fundamental concepts, assessment of routing architectures and protocols, and performance optimization techniques via a systematic survey and synthesis of the related literature. The negative impacts of climate change and the increasing global population on food security and unemployment threats have motivated the adoption of the wireless sensor network (WSN)-based Agri-IoT as an indispensable underlying technology in precision agriculture and greenhouses to improve food production capacities and quality. However, most related Agri-IoT testbed solutions have failed to achieve their performance expectations due to the lack of an in-depth and contextualized reference tutorial that provides a holistic overview of communication technologies, routing architectures, and performance optimization modalities based on users’ expectations. Thus, although IoT applications are founded on a common idea, each use case (e.g., Agri-IoT) varies based on the specific performance and user expectations as well as technological, architectural, and deployment requirements. Likewise, the agricultural setting is a unique and hostile area where conventional IoT technologies do not apply, hence the need for this tutorial. Consequently, this tutorial addresses these via the following contributions: (1) a systematic overview of the fundamental concepts, technologies, and architectural standards of WSN-based Agri-IoT, (2) an evaluation of the technical design requirements of a robust, location-independent, and affordable Agri-IoT, (3) a comprehensive survey of the benchmarking fault-tolerance techniques, communication standards, routing and medium access control (MAC) protocols, and WSN-based Agri-IoT testbed solutions, and (4) an in-depth case study on how to design a self-healing, energy-efficient, affordable, adaptive, stable, autonomous, and cluster-based WSN-specific Agri-IoT from a proposed taxonomy of multi-objective optimization (MOO) metrics that can guarantee an optimized network performance. Furthermore, this tutorial established new taxonomies of faults, architectural layers, and MOO metrics for cluster-based Agri-IoT (CA-IoT) networks and a three-tier objective framework with remedial measures for designing an efficient associated supervisory protocol for cluster-based Agri-IoT networks.

금융 분야의 범주 불균형 문제 해결을 위한 성과 최적화 기반의 부스팅 학습

In this paper, we propose a GMBoost (Geometric Mean-based Boosting) with a direct performance optimization technique to resolve class imbalance problem in financial field, such as bankruptcy, card insolvency, and card fraud. The conventional boosting models including AdaBoost, GBM, and XGBoost are used as benchmarking models for performance comparison. The main findings of 30 rounds of cross validations are as follows. First, the conventional boosting models largely depend on the specificity of majority, but ignore the sensitivity of minority. On the contrary, GMBoost proceeds with balanced learning that simultaneously considers the specificity and the sensitivity at the same time. Second, GMBoost outperforms the conventional boosting algorithms in terms of GM and AUC, and the results of the t-test for GM and AUC also showed that the performance of GMBoost is significantly different from those of the benchmarking models.

Optimization of machine learning approaches for shale gas production forecast

Successful production from shale gas formations has changed the energy outlook of North America. During such process, a significant amount of data related to production and operations has been accumulated, which can be analyzed with machine learning (ML) techniques for production forecasts and performance optimization. In this study, an optimization workflow is proposed for production forecast in shale gas reservoirs, focusing on mitigating the overfitting problem while maximizing its accuracy. More specifically, a new objective function is developed where a penalty term is added to mitigate the model's overfitting tendency. The kernel density estimate (KDE) method guarantees that a similar distribution is shared between the split training and testing subsets. Four machine learning algorithms, including XGBoost, LightGBM, CatBoost, and Random Forest, are utilized and further optimized by the sparrow search algorithm with the newly developed objective function. A stacking approach is finally established based on the optimized algorithm and selected meta-models. Results show that the R2 of all four models has increased to around 0.9 after the optimization process, where XGBoost and LightGBM perform slightly better than the rest. In addition, R2 of the training dataset for XGBoost was decreased from 0.999 to 0.937, while the R2 for the testing dataset was increased from 0.762 to 0.905. This indicates that the overfitting has been mitigated after the optimization with the gap between the training and testing datasets reduced from 0.237 to 0.032. Moreover, stacking the optimized ML algorithms can not further enhance its accuracy. The best performer with SVM as the meta-model can only increase the R2 to 0.910, yet the model complexity increases significantly.

High-Speed DSP Pipelining and Retiming techniques for Distributed-Arithmetic RNS-based FIR Filter Design

Digital FIR Filters plays a major role in many signal processing applications. Generally, these filters are designed with multipliers and adders to find the filter output. This paper acquaints how to reduce the complexity of higher order FIR filter by using performance optimization techniques like retiming and pipelining. The filter’s throughput, energy efficiency, and latency, as well as the complexity of its technology, all need to be improved. By adopting pipelining technique, the arithmetic processes of addition and multiplication are separated. The break addition procedure is retimed. The architecture of Pipelining and Retiming with m-tap filters and n-bit word lengths were designed. The smallest delay achieved by the proposed distributed arithmetic-based FIR Filter with pipelining was 2.564ns for a 4tap implementation receiving an 8bit input, while the largest delay achieved was 56.04ns for a 64-tap implementation receiving a 32-bit word length. Delays as low as 0.68ns for a 4-tap implementation receiving an 8-bit input and as high as 4.53ns for a 64tap implementation receiving a 32bit word length have been achieved by using the suggested distributed arithmetic-based FIR Filter with retiming approach. Delay has been reduced by 73.2% for 4tap with 8bit input and by 91.9% for 64tap with 32bit word length compared to the pipelining approach.

Evaluation of Nanosheet and Forksheet Width Modulation for Digital IC Design in the Sub-3-nm Era

In this article, we provide a comprehensive evaluation of width modulation capabilities of both nanosheet (NS) and forksheet (FS) devices, going from device level to a block level implementation. The main innovation introduced by the FS consists of a dielectric wall added between the p- and nMOS transistors. Leveraging this feature, FS shows approximately the same current behavior as NS, considered a state-of-the-art reference, but reduced parasitic capacitance thanks to its fewer but wider stacked sheets. At block level, an area reduction up to 12% is observed with FS, alongside a 13% power reduction and 10% frequency increase. Following the device comparison, the potential of sheet width modulation as additional power, performance, and area (PPA) optimization technique during synthesis and place and route (PNR) is investigated. A description of the specific steps required to enable this knob in a conventional electronic design automation (EDA) framework is provided. As demonstrated by the obtained experimental results, the same frequency of the single-width implementation can be achieved using mixed libraries with lower power consumption (13% and 16% for NS and FS, respectively), leading to improved energy efficiency. Furthermore, it is shown how designs implemented using FS benefit more from this type of optimization than the ones using NS, with a 12%–15% energy reduction compared to the 8.5%–14% obtained with NS.

A review of machine learning techniques for process and performance optimization in laser beam powder bed fusion additive manufacturing

A review of machine learning techniques for process and performance optimization in laser beam powder bed fusion additive manufacturing

Research on Optimization Design of Fully Parameterized Pump-Jet Propulsion

A pump-jet propulsion system is composed of rotor, stator, and duct. The stator has the front stator type and the rear stator type; the conduit also has the acceleration conduit and the deceleration conduit two forms. It is difficult to design and evaluate the performance of a pump-jet propulsion system because of its complex structure and many changes in component parameters. Due to the limitation of time and cost in the design process of the pump-jet propulsion system, it is difficult to find the optimal scheme in the design space. However, under the guidance of an optimization algorithm, the automatic optimization method can fill the design space with a large number of design schemes. In this paper, the geometry reconstruction technique, hydraulic performance evaluation technique and optimization technique of the pump-jet propulsion system are combined to realize the automation of the whole design process. Firstly, the geometric modeling of the pump-jet propulsion system is completed by a full parametric modeling method, and then the hydrodynamic performance of the pump-jet propulsion system is calculated based on the numerical simulation technique. The radial parameters in the fully parametric configuration of the pump-jet propulsion system were selected as the optimization design variables, and the hydro-dynamic performance was optimized as the objective function. Finally, the pump-jet propulsion system optimization design system was constructed by using the global intelligent optimization algorithm. This study provides a theoretical basis and technical guidance for numerical calculation and configuration optimization design of pump-jet propulsion system.

Deep reinforcement learning in loop fusion problem

Loops’ execution time and resource consumption are one of the interest points and vital issues in the field of appraising complex scientific or computational algorithms. This issue caused the proposal of Loop performance optimization techniques such as fusion. In the literature, loop fusion merges the loops by taking into account a set of properties associated with the loops or the system on which the resulting code will be executed. The number of these factors and their interactions on the one hand, and the high runtime of available comprehensive approaches, on the other hand, reveals the need for a new method that could be concerned for further progress in solving this NP-hard problem. For the first time, Deep Reinforcement Learning Loop Fusion (DRLLF) advanced to be an ideal solution for the challenge in this article. For the proposed framework, a particular matrix is configured as the inputs of a deep neural network based on the information of the problem, namely data dependencies, data reuse, loops’ types, and computer system’s register size. These randomly generated matrixes are used in the training phase by reinforcement learning to get the imperative experience on predicting a profitable distribution over loops’ various fusion orders. In the evaluations performed, the presented algorithm was able to achieve the same or better performance in terms of speedup rate, comparing with the methods under study, approximately averaged in 7.36 percent better results. The considerable improvement observed in the results, besides the low run time, proves the comprehensiveness and superiority of this approach.

A Survey of Performance Tuning Techniques and Tools for Parallel Applications

Automatic parallelization of sequential programs combined with auto-tuning is an alternative to manual parallelization. With wider research directions and the increased number of performance tuning tools that have been developed, it has become increasingly difficult to choose a particular tuning tool. This paper reviews the fundamentals of different performance optimization and tuning techniques. It also surveys several tuning frameworks and classifies them into different groups based on their criteria. Developing benchmarks for HPC and validating their accuracy are demanding tasks for computer architects, researchers, and application developers. In addition to providing a survey of performance-tuning tools, we also performed a detailed review of current benchmarks and discussed the requirements for future benchmarks. We performed a detailed comparison of these tuning tools based on other features such as speedup and infrastructure details. We believe that this study will be a very useful resource for parallel computing communities, especially for early-stage parallel computing and performance researchers to gain exposure to existing performance optimization options.

Computation Performance Optimization Technique of Shortest Path Routing Algorithm in Networks Using Out-degree

Objective: This paper presents a novel technique in Dijkstra's routing algorithm by considering the concept of out degree which will decrease the computational cost and increase the speed of execution.Methods: In this proposed method the time taken by this algorithm that runs as a combination of the input distance is calculated. Conversely, a computation time complexity measures how efficient it is. accuracy of this method. The Shortest Path Problem is the problem of finding a path in a digraph between two vertices or nodes to minimize the sum of the weights of the edges of its constituent. In the existing approach it specifies a number of nodes as the plot unfolds and calculates the quickest route between it and most other nodes. which will complex the speed of execution.Findings: The proposed out degree approach will resolve that drawback by computing two minimum distance from initial table and this operates from the source node and determines the shortest path over the network as a whole. It also guarantees to find a globally optimal solution path with accurate results.Novelty: The algorithm’s description and comparison are presented in graphical ways to determine the algorithm’s features. The analysis shows that the best route from source to destination will be established which provides the shortest distance. The optimal pathways discovered as a result of the analysis which reduce the distance traveled by the company in shipping products and reduce the time and cost of delivery.

Urban heat-mitigating building form and façade framework

The urban heat island (UHI) effect is one of the most critical issues for dense urban environments; building forms and facades increase this effect and building energy use intensity (EUI). This research hypothesizes that by using a design-based approach, a new generation of climate-responsive building forms and facades can be developed that balance the outdoor and indoor environments and minimize the UHI effect. A novel design workflow is proposed to integrate UHI mitigation strategies, urban microclimate simulation methods, and building performance optimization techniques during the early stages of the architectural design process, employing elements such as parametric modeling and objective optimization. As a proof of concept, redesigning a 56-storey core and shell office tower in downtown Chicago, IL is found to improve outdoor thermal comfort by up to 1.26°C, EUI by 0.79 kWh/m2, and unusable daylight by 50.90%, as compared to the existing design.

A Survey of Machine Learning-Based System Performance Optimization Techniques

Recently, the machine learning research trend expands to the system performance optimization field, where it has still been proposed by researchers based on their intuitions and heuristics. Compared to conventional major machine learning research areas such as image or speech recognition, machine learning-based system performance optimization fields are at the beginning stage. However, recent papers show that this approach is promising and has significant potential. This paper reviews 11 machine learning-based system performance optimization approaches from nine recent papers based on well-known machine learning models such as perceptron, LSTM, and RNN. This survey provides a detailed design and summarizes model, input, output, and prediction method of each approach. This paper covers various system performance areas from the data structure to essential system components of a computer system such as index structure, branch predictor, sort, and cache management. The result shows that machine learning-based system performance optimization has an important potential for future research. We expect that this paper shows a wide range of applicability of machine learning technology and provides a new perspective for system performance optimization.

Multi-objective optimization of building energy performance and indoor thermal comfort by combining artificial neural networks and metaheuristic algorithms

During the last few years, multi-objective optimization processes have become one of the main challenges for energy efficiency in buildings. In this work, a new efficient multi-objective optimization method, based on the Building Performance Optimization (BPO) technique, has been developed to improve the indoor thermal comfort and energy performance of residential buildings, i.e. a Moroccan ground floor + first floor (GFFF) house located in Marrakech region (5th climatic zone according to the Thermal Building Code in Morocco). The most influential design variables have been well explored in order to find the optimal trade-off between these two objectives. Indeed, this technique is based on the integration of Artificial Neural Networks (ANNs), in particular Multilayer Feedforward Neural Networks (MFNN), coupled with the most commonly used metaheuristic algorithms, i.e. Non-dominated Sorting Genetic Algorithm (NSGA-II), Multi-Objective Particle Swarm Optimization (MOPSO) and Multi-Objective Genetic Algorithm (MOGA), in order to minimize computation time as much as possible. The TRNSYS software was used to establish the various dynamic thermal simulations required to create the database, from which the ANNs were able to set up their learning. The results show that this methodology is being used successfully, leading to different proposed solutions in terms of building envelope design. However, only the solutions using MOPSO are finally retained, as they have shown the greatest desired performance compared to the others. Thus, the thermal needs, particularly those for heating and cooling, have been significantly reduced to 74.52% of the total, while improving the indoor thermal comfort by 4.32% compared to the base design. Finally, we strongly recommend this methodology to the different actors in this field, including designers, engineers, architects, engineering offices, etc., when several objectives need to be contrasted while simultaneously considering several design variables.

GPU-Accelerated Flexible Molecular Docking.

Virtual screening is a key enabler of computational drug discovery and requires accurate and efficient structure-based molecular docking. In this work, we develop algorithms and software building blocks for molecular docking that can take advantage of graphics processing units (GPUs). Specifically, we focus on MedusaDock, a flexible protein-small molecule docking approach and platform. We accelerate the performance of the coarse docking phase of MedusaDock, as this step constitutes nearly 70% of total running time in typical use-cases. We perform a comprehensive evaluation of the quality and performance with single-GPU and multi-GPU acceleration using a data set of 3875 protein-ligand complexes. The algorithmic ideas, data structure design choices, and performance optimization techniques shed light on GPU acceleration of other structure-based molecular docking software tools.