Common Optimization Methods Research Articles

Multi-Layer Energy Management and Strategy Learning for Microgrids: A Proximal Policy Optimization Approach

An efficient energy management system (EMS) enhances microgrid performance in terms of stability, safety, and economy. Traditional centralized or decentralized energy management systems are unable to meet the increasing demands for autonomous decision-making, privacy protection, global optimization, and rapid collaboration simultaneously. This paper proposes a hierarchical multi-layer EMS for microgrid, comprising supply layer, demand layer, and neutral scheduling layer. Additionally, common mathematical optimization methods struggle with microgrid scheduling decision problem due to challenges in mechanism modeling, supply–demand uncertainty, and high real-time and autonomy requirements. Therefore, an improved proximal policy optimization (PPO) approach is proposed for the multi-layer EMS. Specifically, in the centrally managed supply layer, a centralized PPO algorithm is utilized to determine the optimal power generation strategy. In the decentralized demand layer, an auction market is established, and multi-agent proximal policy optimization (MAPPO) algorithm with an action-guidance-based mechanism is employed for each consumer, to implement individual auction strategy. The neutral scheduling layer interacts with other layers, manages information, and protects participant privacy. Numerical results validate the effectiveness of the proposed multi-layer EMS framework and the PPO-based optimization methods.

A model-based brain switch via periodic motor imagery modulation for asynchronous brain-computer interfaces

Objective. Brain switches provide a tangible solution to asynchronized brain-computer interface, which decodes user intention without a pre-programmed structure. However, most brain switches based on electroencephalography signals have high false positive rates (FPRs), resulting in less practicality. This research aims to improve the operating mode and usability of the brain switch. Approach. Here, we propose a novel virtual physical model-based brain switch that leverages periodic active modulation. An optimization problem of minimizing the triggering time subject to a required FPR is formulated, numerical and analytical approximate solutions are obtained based on the model. Main results. Our motor imagery (MI)-based brain switch can reach 0.8FP/h FPR with a median triggering time of 58 s. We evaluated the proposed brain switch during online device control, and their average FPRs substantially outperformed the conventional brain switches in the literature. We further improved the proposed brain switch with the Common Spatial Pattern (CSP) and optimization method. An average FPR of 0.3 FPs/h was obtained for the MI-CSP-based brain switch, and the average triggering time improved to 21.6 s. Significance. This study provides a new approach that could significantly reduce the brain switch’s FPR to less than 1 Fps/h, which was less than 10% of the FPR (decreasing by more than a magnitude of order) by other endogenous methods, and the reaction time was comparable to the state-of-the-art approaches. This represents a significant advancement over the current non-invasive asynchronous BCI and will open widespread avenues for translating BCI towards clinical applications.

An improvement on common optimization methods based on SuperstarGAN

Image processing has long been a focal point of research, offering avenues to enhance image clarity and transfer image features. Over the past decade, Generative Adversarial Networks (GANs) have played a pivotal role in the field of image conversion. This study delves into the world of GANs, focusing on the SuperstarGAN model and its optimization techniques. SuperstarGAN, an evolution of the well-known StarGAN, excels in multi-domain image-to-image conversion, overcoming limitations and offering versatility. To better understand its optimization, this study explored the effects of different optimizers, such as Adam, SGD, and Nadam, on SuperstarGAN's performance. Using the CelebA Face Dataset with 200 million images and 40 features, this study conducted experiments to compare these optimizers. The results revealed that while SGD and Nadam can achieve comparable results to Adam, they require more iterations and careful tuning, with SGD showing slower convergence. Nadam, with its oscillatory nature, shows promise but requires proper learning rate adjustments. This research sheds light on the critical role of optimizer choice in training SuperstarGAN. Adam emerges as the most efficient and stable option, but further exploration of Nadam's potential is warranted. This study contributes to advancing the understanding of optimization techniques for generative adversarial networks, with implications for high-quality facial image generation and beyond.

Improving efficiency of parallel inverters operation in island mode microgrids

DC/AC inverters play a vital role in microgrids, efficiently converting renewable energy into usable AC power. Parallel operation of inverters presented numerous challenges, including maximizing system efficiency, minimizing circulating current, and maximizing system accuracy. This proposal introduces an analytical optimization technique designed to enhance the efficiency of paralleled inverters in microgrid systems while minimizing circulating current. The system parameter estimation is performed with a rapid recursive least squares (RLS) estimator. An optimized proportional-integral-derivative (PID) controller achieves high accuracy and streamlining system construction. The performance of the proposed optimizer is compared to common optimization methods, such as particle swarm optimization (PSO), neural networks, interior search, and interior point optimizers, focusing on system efficiency and eliminating circulating currents. Simulation investigations validated the method's applicability and demonstrated the proposed optimizer's superiority in efficiency, stability, and limiting circulating currents. It also achieved zero execution time, significantly outperforming alternatives like the neural network optimizer, which took 0.693 s. In various scenarios, the proposed optimizer improved system efficiency by 3% compared to the equally shared current system.

Void fraction measurement of gas–liquid two-phase flow based on compressed sensing theory and capacitively coupled electrical resistance tomography system

Capacitively coupled electrical resistance tomography (CCERT) is an innovative technique for electrical resistance tomography based on capacitively coupled contactless conductivity detection (C4D). Despite its potential, there are only a few studies on image reconstruction algorithms for CCERT. To address this, a CCERT measurement system is developed, and the compressed sensing (CS) theory is applied to the inverse problem imaging of CCERT to improve the image reconstruction quality and speed. Firstly, a mathematical CCERT image reconstruction model is constructed under CS theory and three algorithms under CS theory are employed to solve the convex optimization problem of the reconstruction model. Then a thresholding operation is applied to obtain the post-processed image and compared it with the classical linear back projection and Landweber algorithms. The simulation results demonstrate that the Barzilai–Borwein gradient projection for sparse reconstruction (GPSR-BB) algorithm yields more satisfactory imaging results than the other four algorithms. Additionally, three sensitivity matrix optimization methods for GPSR-BB algorithm are compared and find that the new method of screening the rows of the sensitivity matrix to zero is more effective than the other two common optimization methods in terms of reconstructed image quality. Finally, the static experiments of void fraction measurement are conducted using the developed CCERT system. The results indicated that the absolute error of void fraction measurement by GPSR-BB algorithm was less than 6.59% in the range of void fraction from 0.90% to 66.29%.

Process optimization of chain-die forming for asymmetric channels by an image-based machine learning method

The coupled springback of twist and angle is a major shape error for high-strength asymmetric channels in the automotive industry. In the process design stage of sheet metal forming, scalar-based surrogate models such as artificial neural networks, are the common optimization method to predict springback and reduce shape error. However, such models fail to learn and utilize the location information of the formed part, which will result in poor prediction accuracy and compensation effect for the coupled springback in asymmetric channels. Therefore, in this paper, based on the U-net convolutional neural network (CNN), an image-based surrogate model is established to predict the dimensional deviation full field of the formed part. An attention-based ResNet layer is proposed and replaced with the bottleneck of the U-net-style base network to enhance the sensitivity of the network for different regions of die surface in the sheet metal forming. An automatic data preprocessing method is developed to convert the design parameters and simulation results into unified image data. After training, the proposed CNN surrogate model shows a highly accurate prediction for the dimensional deviation full field in real-time. Furthermore, an image-based optimization architecture, in conjunction with a differential evolution algorithm with self-adaptive factors, the automatic data preprocessing method, and the CNN surrogate model, is developed to minimize the dimensional deviations over the entire region of the formed part. Finally, the validation experiments of chain-die forming are carried out by using optimized design parameters and the forming accuracy has been significantly improved.

A novel metaheuristic MPPT technique based on enhanced autonomous group Particle Swarm Optimization Algorithm to track the GMPP under partial shading conditions - Experimental validation

In order to maximize the benefit of solar panels, it is crucial to enhance the operating efficiency of PV systems. Usually, the PV arrays are exposed to harsh outdoor operating conditions that affect their operating efficiency. Extraction of global maximum power point (GMPP) among the other local MPPs (LMPPs) from the distorted characteristics of the PV array is considered the main challenge of PV system controllers under partial shading conditions (PSCs). Classical MPP tracking (MPPT) techniques have incompetent performance in tracing the GMPP under the PSCs of PV arrays. Therefore, many metaheuristics-based MPPT techniques are developed by researchers to defeat the problems related to the PSCs and cope with GMPP and other LMPPs. Among all metaheuristic algorithms, Particle Swarm Optimization (PSO) and its modified versions are the most common optimization methods owing to their easy implementation, simplicity, and low computational burden. On the other hand, the standard PSO (SPSO) algorithm suffers from low convergence speed, and difficulty in tuning its parameters, and it may be trapped in LMPP under fast changes in solar radiation and/or PSCs. In order to tackle these issues, a novel metaheuristic MPPT technique based on an Enhanced Autonomous Group PSO (EAGPSO) algorithm is proposed. A comparative study between EAGPSO and other versions of PSO techniques is performed to emphasize the efficacy of the EAGPSO technique. The simulation results demonstrate that the proposed EAGPSO algorithm is superior to the counterpart versions of the PSO algorithms in terms of fast-tracking time, high obtained power, and high energy efficiency. Experimental validation of the EAGPSO technique has been conducted using an array of 3*3 series–parallel PV panels under PSC. The EAGPSO algorithm results have been compared with SPSO and AGPSO techniques. The experimental results show that, among the investigated MPPT techniques, the proposed EAGPSO technique has the highest MPPT efficiency of 98.6% and the lowest power oscillations of about 2.03% of the tracked power with a tracking time of 2.6 s.

Response Surface Methodology in The Optimization of Natural Product Extraction and Separation Techniques

This paper focuses on the research application of response surface methodology in the optimization of natural product extraction and separation technology, and reviews the common experimental design optimization methods, the basic concepts, classification and research methods of response surface methodology, as well as the research progress in the optimization of anthocyanin and flavonoids plant extraction; analyzes the research prospects of response surface methodology, and predicts the research hotspots and directions in the light of actual practice.

Monitoring variability in parameter estimates for lumped parameter models of the systemic circulation using longitudinal hemodynamic measurements

BackgroundPhysics-based cardiovascular models are only recently being considered for disease diagnosis or prognosis in clinical settings. These models depend on parameters representing the physical and physiological properties of the modeled system. Personalizing these parameters may give insight into the specific state of the individual and etiology of disease. We applied a relatively fast model optimization scheme based on common local optimization methods to two model formulations of the left ventricle and systemic circulation. One closed-loop model and one open-loop model were applied. Intermittently collected hemodynamic data from an exercise motivation study were used to personalize these models for data from 25 participants. The hemodynamic data were collected for each participant at the start, middle and end of the trial. We constructed two data sets for the participants, both consisting of systolic and diastolic brachial pressure, stroke volume, and left-ventricular outflow tract velocity traces paired with either the finger arterial pressure waveform or the carotid pressure waveform.ResultsWe examined the feasibility of separating parameter estimates for the individual from population estimates by assessing the variability of estimates using the interquartile range. We found that the estimated parameter values were similar for the two model formulations, but that the systemic arterial compliance was significantly different (p < 10^{-6}) depending on choice of pressure waveform. The estimates of systemic arterial compliance were on average higher when using the finger artery pressure waveform as compared to the carotid waveform.ConclusionsWe found that for the majority of participants, the variability of parameter estimates for a given participant on any measurement day was lower than the variability both across all measurement days combined for one participant, and for the population. This indicates that it is possible to identify individuals from the population, and that we can distinguish different measurement days for the individual participant by parameter values using the presented optimization method.

Topology Optimization Design of Multi-Input-Multi-Output Compliant Mechanisms with Hinge-Free Characteristic and Totally Decoupled Kinematics

A new multi-constraint optimization model with the weighted objective function is proposed to design the multi-input-multi-output (MIMO) compliant mechanisms. The main feature of this work is that both the two notable problems related to the de facto hinge and the movement coupling are tackled simultaneously in the topological synthesis of MIMO compliant mechanisms. To be specific, the first problem is the severe stress concentration in the flexible hinge areas, and it is solved by the introduction of input and output compliances into the objective function, which could facilitate the optimization to strike a good balance between structural flexibility and stiffness. The second problem is the high degree of control complexity caused by the coupled outputs and inputs, and it is addressed by achieving the complete decoupling with two groups of extra constraints that are used to suppress the input coupling and the output coupling, respectively. As the most common and effective topology optimization method, the Solid Isotropic Material with Penalization (SIMP)-based density method is adopted here to obtain the optimized configurations. After the analytical sensitivity deduction related to the weighted objective function and constraints, two typical numerical examples are presented to demonstrate the validity of the proposed topology optimization framework in designing the hinge-free and completely decoupled MIMO compliant mechanisms.

An adaptive fuzzy logic control technique for LVRT enhancement of a grid-integrated DFIG-based wind energy conversion system

The fast improvement of wind energy conversion technology with optimization algorithms has recently received a lot of attention. However, their slow convergence speed, huge computational load, and low efficiency are the main drawbacks and to improve these disadvantages, a new adaptive fuzzy logic controller strategy is proposed to enhance the low voltage ride-through of the grid-connected doubly fed induction generator during severe grid faults. The rotor side converter and the grid side converter are controlled using an adaptive fuzzy logic controller topology under cascaded structure to improve the performance. The novel application of the ‘generalized variable step-size diffusion continuous mixed p-norm’ adaptive filtering algorithm is proposed to modify the calibrating factors of the fuzzy logic controllers at a rapid convergence speed with low normalized misalignment error. The proposed adaptive algorithm-based fuzzy logic controller has a better low voltage ride-through improvement capability than that of using the particle swarm optimization-based proportional-integral controller during severe grid disturbances. The convergence speed of the proposed adaptive filtering algorithm is compared to that of existing algorithms such as least mean fourth, least mean square, and continue mixed p-norm algorithms. In addition, a comparison is also made with different common optimization methods. The field-programmable gate array-based real-time experimental results are presented to validate the proposed adaptive control topology.

A SCIENCE MAPPING APPROACH BASED REVIEW OF MODEL PREDICTIVE CONTROL FOR SMART BUILDING OPERATION MANAGEMENT

Model predictive control (MPC) for smart building operation management has become an increasingly popular and important topic in the academic community. Based on a total of 202 journal articles extracted from Web of Science, this study adopted a science mapping approach to conduct a holistic review of the literature sample. Chronological trends, contributive journal sources, active scholars, influential documents, and frequent keywords of the literature sample were identified and analyzed using science mapping. Qualitative discussions were also conducted explore in details the objectives and data requirements of MPC implementation, different modeling approaches, common optimization methods, and associated model constraints. Three research gaps and future directions of MPC were presented: the selection and establishment of MPC central model, the capability and security of processing massive data, and the involvement of human factors. This study provides a big picture of existing research on MPC for smart building operations and presents findings that can serve as comprehensive guides for researchers and practitioners to connect current research with future trends.

A brief review of variable stiffness and damping magnetorheological fluid dampers

With the development of magnetorheological fluid (MRF) technology, variable stiffness and damping MRF dampers as semi-active controllers gradually show their unique advantages in the vibration control field. This paper reviews recent journal articles on the principles, characteristics, structure design, modelling and applications of variable stiffness and damping MRF dampers. The principle includes the force behaviour of the MRF fluid and the mechanism of the variable stiffness and damping MRF damper. Characteristics introduce the damper’s force-displacement relationships, which have frequency dependence and amplitude dependence. Following this, typical dampers structures are discussed, including electromagnetic design and coil configuration. After that, some common models and parameter optimization methods are presented. Furthermore, taking vehicle suspension as an example, the applications of variable stiffness and damping MRF dampers in some fields are proposed in this paper.

An Efficient Heap Based Optimizer Algorithm for Feature Selection

The heap-based optimizer (HBO) is an innovative meta-heuristic inspired by human social behavior. In this research, binary adaptations of the heap-based optimizer B_HBO are presented and used to determine the optimal features for classifications in wrapping form. In addition, HBO balances exploration and exploitation by employing self-adaptive parameters that can adaptively search the solution domain for the optimal solution. In the feature selection domain, the presented algorithms for the binary Heap-based optimizer B_HBO are used to find feature subsets that maximize classification performance while lowering the number of selected features. The textitk-nearest neighbor (textitk-NN) classifier ensures that the selected features are significant. The new binary methods are compared to eight common optimization methods recently employed in this field, including Ant Lion Optimization (ALO), Archimedes Optimization Algorithm (AOA), Backtracking Search Algorithm (BSA), Crow Search Algorithm (CSA), Levy flight distribution (LFD), Particle Swarm Optimization (PSO), Slime Mold Algorithm (SMA), and Tree Seed Algorithm (TSA) in terms of fitness, accuracy, precision, sensitivity, F-score, the number of selected features, and statistical tests. Twenty datasets from the UCI repository are evaluated and compared using a set of evaluation indicators. The non-parametric Wilcoxon rank-sum test was used to determine whether the proposed algorithms’ results varied statistically significantly from those of the other compared methods. The comparison analysis demonstrates that B_HBO is superior or equivalent to the other algorithms used in the literature.

Numerical simulation and sensitivity analysis of automatic optimization of gas extraction well locations in abandoned mines

There are large numbers of abandoned mines in China and abroad, among of which stores relative abundant coal-bed gas. The gas extraction from abandoned mines can be used for power generation or industrial gas utilization. In this paper, the present situation of abandoned mines in China and abroad was studied, and the disadvantages of the common layout and optimization methods of gas extraction in abandoned mines were pointed out. Based on the analysing of reservoir conditions and physical property parameters, the finite element software COMSOL was used to do the numerical simulation of extraction well automatically optimize, and take shenbei coal mine as an example. Moreover, an automatic optimization APP of well location was developed based on the results of simulation. Finally, the parameter sensitivity of gas extraction is analysed. Through the study, it is found that the most basic and common method of gas extraction in abandoned mines is empirical method, which lacks of theoretic support and accuracy, thus causing unreasonable extraction and waste of resources. In this study, the COMSOL numerical simulation software was used to explore the numerical simulation method of automatic well position optimization combined with the optimization algorithm. The Nelder-Mead algorithm was selected as the optimization algorithm after comparing the advantages and disadvantages of various algorithms. The operation flow and type of Nelder-Mead algorithm were analysed, and the optimization parameters of numerical simulation were set. Through the numerical simulation, the function of automatically searching the high-gas production position near the preset well position was realized, and the automatic optimization APP for gas extraction in abandoned mines was developed. Finally, the sensitivity of parameters to gas production in different well locations under three different pumping conditions which are horizontal, vertical and bottom-hole negative pressure was analysed; and the simulation results were discussed.

Alpert multi-wavelets for functional inverse problems: direct optimization and deep learning

Computational engineering models often contain unknown entities (e.g. parameters, initial and boundary conditions) that require estimation from other measured observable data. Estimating such unknown entities is challenging when they involve spatio-temporal fields because such functional variables often require an infinite-dimensional representation. We address this problem by transforming an unknown functional field using Alpert wavelet bases and truncating the resulting spectrum. Hence the problem reduces to the estimation of few coefficients that can be performed using common optimization methods. We apply this method on a one-dimensional heat transfer problem where we estimate the heat source field varying in both time and space. The observable data is comprised of temperature measured at several thermocouples in the domain. This latter is composed of either copper or stainless steel. The optimization using our method based on wavelets is able to estimate the heat source with an error between 5% and 7%. We analyze the effect of the domain material and number of thermocouples as well as the sensitivity to the initial guess of the heat source. Finally, we estimate the unknown heat source using a different approach based on deep learning techniques where we consider the input and output of a multi-layer perceptron in wavelet form. We find that this deep learning approach is more accurate than the optimization approach with errors below 4%.

On the application of the LET-model for the approximation of core relative phase permeabilities

The article deals with the problem of adequate interpretation of core relative phase permeabilities and the use of the LET-model in application to this problem. Some existing analytical dependencies for the interpretation of laboratory studies of the OFP are briefly analyzed: the Corey model, the Sigmund and McCaffery model, the Chierici model, the LET-model. The advantage of the LET-model and the high evaluation of this model in independent studies of the authors are described. In the process of preliminary analysis, the authors of the article revealed that the non-trivial issue of determining the adjustable coefficients of the model using mathematical methods was not sufficiently covered, which became the central object of the work. Further, the application of the LET-model is formalized and the methodology for determining the adjustable coefficients of the model (L, E and T) is described. The task of selecting model parameters is interpreted as a multidimensional optimization problem. Also, a class of nonlinear least squares methods was defined to solve the problem. Several common numerical optimization methods for selecting the coefficients of the model are considered: the brute-force search method, the Levenberg-Marquardt algorithm method, the trust region method. Computational experiments were carried out to assess their applicability and comparative analysis of optimization algorithms. Based on the results of the experiments, it was concluded that the method of confidence domains is suitable for practical application, since they converge to the optimal solution quickly enough and demonstrates high accuracy in the selection of coefficients L, E, T. The final part of the article demonstrates the difference between the LET-model and the basic Corey model when approximating core relative phase permeabilities.

Optimization of injection molding process parameters based on GA-ELM-GA

The most common optimization method for the optimization of injection mold process parameters is range analysis, but there is often a nonlinear coupling relationship between injection molding process parameters and quality indicators. Therefore, it is difficult to find the optimal process combination in range analysis. In this article, a genetic algorithm optimized extreme learning machine network model (GA-ELM) combined with genetic algorithm (GA) was proposed to optimize the process parameters of the injection mold. Take the injection molding process parameter optimization of an electrical appliance buckle cover shell as an example. In order to find the process parameters corresponding to the minimum warpage deformation, an orthogonal experiment was designed and the results of the orthogonal experiment were analyzed. Then, the corresponding optimal process combination and the degree of influence of process parameters on the warpage deformation were obtained. At the same time, the extreme learning machine network model (GA-ELM) optimized by the genetic algorithm was used to predict the warpage deformation of the plastic part. The trained GA-ELM model can map non-linear coupling relationship between the five process parameters and the warpage deformation well. And the optimal process parameters in the trained GA-ELM network model was searched by the powerful optimization ability of genetic algorithm. Generally speaking, the warpage deformation after optimization by range analysis is reduced by 6.7% compared with the minimum warpage after optimization by orthogonal experiment. But compared to the minimum warpage deformation after orthogonal experiment optimization, the warpage deformation after GAELM-GA optimization is reduced by 22%, which is better than that of the range analysis, thus verifying the feasibility and the optimization of the optimization method. This optimization method provides a certain theoretical reference and technical support for the field involving the optimization of process parameters.

Selecting process parameters of crack-free Ni60A alloy coating prepared by coaxial laser cladding

Ni60A alloy coating prepared by laser cladding is limited by the high cracking susceptibility. In this study, the crack-free Ni60A coating is obtained by selecting suitable process parameters. In order to avoid size effect on heat dissipation and stress evolution, the width and height models of cladding track are established to obtain same size track with different process parameters. These two models divide the three main parameters laser power, powder feed rate and scanning speed into two groups, namely line energy (the ratio of laser power to scanning speed) and powder feed rate, and determine their experimental values. Then the influence of determined line energy and powder feed rate on cracking are tested, which significantly reduces the amount of tests compared with common optimization methods. The results show that large line energy and low powder feed rate are helpful to prevent crack, but lead to low microhardness. In order to obtain crack-free Ni60A coating with higher microhardness, the combination method of preheating to 300 °C and adding insulating plate under the substrate is further studied, which can not only reduce thermal damage to the substrate but also improve the microhardness up to 510–553 HV.

Optimization of Data Quality Related EMG Feature Extraction Parameters to Increase Hand Movement Classification Accuracy.

Many biomedical robotic interfaces (e.g., prostheses, exoskeletons) classify or estimate user movement intent based on features extracted from measured electromyograms (EMG). In most cases, the parameters of feature extraction are determined heuristically or assigned arbitrary values. We propose a more rigorous method, numerical optimization, to systematically identify parameters that maximize classification accuracy based on EMG signal characteristics. In this study, we used simulated annealing, a common global numerical optimization method, to find the optimal values of three feature extraction parameters based on the root mean square (rms) magnitude of the EMG signal. The EMG data, obtained from a public database, had been measured from 2 muscles (one hand flexor and one hand extensor) of 5 able-bodied participants performing 6 different movement tasks. Using optimization, we increased the offline movement classification accuracy by 3-5% for each participant and from 79.91% to 92.25% overall. The value of one optimized parameter (threshold of Wilson amplitude) was strongly correlated with the rms magnitude of the EMG signal (R2=0.81). Other parameters were suspected to be related to signal noise, since no strong correlation with rms magnitude was observed. Future studies will refine the optimization approach and test its practicality and effectiveness for improving online classification accuracy with robotic interfaces.