Comparison Of Algorithms Research Articles

Partial prior transfer learning based on self-attention CNN for EEG decoding in stroke patients.

The utilization of motor imagery-based brain-computer interfaces (MI-BCI) has been shown to assist stroke patients activate motor regions in the brain. In particular, the brain regions activated by unilateral upper limb multi-task are more extensive, which is more beneficial for rehabilitation, but it also increases the difficulty of decoding. In this paper, self-attention convolutional neural network based partial prior transfer learning (SACNN-PPTL) is proposed to improve the classification performance of patients' MI multi-task. The backbone network of the algorithm is SACNN, which accords with the inherent features of electroencephalogram (EEG) and contains the temporal feature module, the spatial feature module and the feature generalization module. In addition, PPTL is introduced to transfer part of the target domain while preserving the generalization of the base model while improving the specificity of the target domain. In the experiment, five backbone networks and three training modes are selected as comparison algorithms. The experimental results show that SACNN-PPTL had a classification accuracy of 55.4%±0.17 in four types of MI tasks for 22 patients, which is significantly higher than comparison algorithms (P < 0.05). SACNN-PPTL effectively improves the decoding performance of MI tasks and promotes the development of BCI-based rehabilitation for unilateral upper limb.

A novel sharding algorithm based on TOPSIS method for blockchain systems

PurposePerformance optimization algorithms based on node attributes are of great importance for sharding blockchain systems. Currently, existing studies on blockchain sharding algorithms consider only random selection sharding strategies. However, the random selection strategy does not perfectly utilize the performance of a node to break the bottleneck of blockchain performance.Design/methodology/approachThis paper proposes a blockchain sharding algorithm called TOPSIS Optimization Sharding System (TOSS), which is based on entropy weight method, relative Euclidean distance and the Technique for Order Preference by Similarity to Ideal Solution (TOPSIS). It defines a multi-attribute matrix to assess node performance and applies TOPSIS for scoring nodes. Then, an algorithm based on the TOPSIS method is proposed to calculate the performance score of each data node. In addition, an entropy weighting method is introduced to obtain the weights of each attribute to balance the impact of dimensional differences of attributes on the attribute weights. Nodes are ranked by composite scores to guide partitioning.FindingsThe effectiveness of the proposed algorithm in this paper is verified by comparing it with various comparative algorithms. The experimental results show that the TOSS algorithm outperforms the comparison algorithms in terms of performance improvement for the blockchain system, and the throughput metrics are improved by about 20% in comparison.Originality/valueThis study introduces a novel approach to blockchain sharding by incorporating the entropy weight method and relative Euclidean distance TOPSIS into the sharding process. This approach allows for a more effective utilization of node performance attributes, leading to significant improvements in system throughput and overall performance, addressing the limitations of the random selection strategy commonly used in existing algorithms.

Unmanned Aerial Vehicle Photogrammetry for Monitoring the Geometric Changes of Reclaimed Landfills

Monitoring reclaimed landfills is essential for ensuring their stability and monitoring the regularity of facility settlement. Insufficient recognition of the magnitude and directions of these changes can lead to serious damage to the body of the landfill (landslides, sinkholes) and, consequently, threaten the environment and the life and health of people near landfills. This study focuses on using UAV photogrammetry to monitor geometric changes in reclaimed landfills. This approach highlights the advantages of UAVs in expanding the monitoring and providing precise information critical for decision-making in the reclamation process. This study presents the result of annual photogrammetry measurements at the Słabomierz–Krzyżówka reclaimed landfill, located in the central part of Poland. The Multiscale Model to Model Cloud Comparison (M3C2) algorithm was used to determine deformation at the landfill. The results were simultaneously compared with the landfill’s reference (angular–linear) measurements. The mean vertical displacement error determined by the photogrammetric method was ±2.3 cm. The results showed that, with an appropriate measurement methodology, it is possible to decide on changes in geometry reliably. The collected 3D data also gives the possibility to improve the decision-making process related to repairing damage or determining the reclamation direction of the landfill, as well as preparing further development plans.

3D Point Cloud Fusion Method Based on EMD Auto-Evolution and Local Parametric Network

Although the development of high-resolution remote sensing satellite technology has made it possible to reconstruct the 3D structure of object-level features using satellite imagery, the results from a single reconstruction are often insufficient to comprehensively describe the 3D structure of the target. Therefore, developing an effective 3D point cloud fusion method can fully utilize information from multiple observations to improve the accuracy of 3D reconstruction. To this end, this paper addresses the problems of shape distortion and sparse point cloud density in existing 3D point cloud fusion methods by proposing a 3D point cloud fusion method based on Earth mover’s distance (EMD) auto-evolution and local parameterization network. Our method is divided into two stages. In the first stage, EMD is introduced as a key metric for evaluating the fusion results, and a point cloud fusion method based on EMD auto-evolution is constructed. The method uses an alternating iterative technique to sequentially update the variables and produce an initial fusion result. The second stage focuses on point cloud optimization by constructing a local parameterization network for the point cloud, mapping the upsampled point cloud in the 2D parameter domain back to the 3D space to complete the optimization. Through these two steps, the method achieves the fusion of two sets of non-uniform point cloud data obtained from satellite stereo images into a single, denser 3D point cloud that more closely resembles the true target shape. Experimental results demonstrate that our fusion method outperforms other classical comparison algorithms for targets such as buildings, planes, and ships, and achieves a fused RMSE of approximately 2 m and an EMD accuracy better than 0.5.

Optimizing Microservice Deployment in Edge Computing with Large Language Models: Integrating Retrieval Augmented Generation and Chain of Thought Techniques

Large Language Models (LLMs) have demonstrated impressive capabilities in autogenerating code based on natural language instructions provided by humans. We observed that in the microservice models of edge computing, the problem of deployment latency optimization can be transformed into an NP-hard mathematical optimization problem. However, in the real world, deployment strategies at the edge often require immediate updates, while human-engineered code tends to be lagging. To bridge this gap, we innovatively integrated LLMs into the decision-making process for microservice deployment. Initially, we constructed a private Retrieval Augmented Generation (RAG) database containing prior knowledge. Subsequently, we employed meticulously designed step-by-step inductive instructions and used the chain of thought (CoT) technique to enable the LLM to learn, reason, reflect, and regenerate. We decomposed the microservice deployment latency optimization problem into a collection of granular sub-problems (described in natural language), progressively providing instructions to the fine-tuned LLM to generate corresponding code blocks. The generated code blocks underwent integration and consistency assessment. Additionally, we prompted the LLM to generate code without the use of the RAG database for comparative analysis. We executed the aforementioned code and comparison algorithm under identical operational environments and simulation parameters, conducting rigorous result analysis. Our fine-tuned model significantly reduced latencies by 22.8% in handling surges in request flows, 37.8% in managing complex microservice types, and 39.5% in processing increased network nodes compared to traditional algorithms. Moreover, our approach demonstrated marked improvements in latency performance over LLMs not utilizing RAG technology and reinforcement learning algorithms reported in other literature. The use of LLMs also highlights the concept of symmetry, as the symmetrical structure of input-output relationships in microservice deployment models aligns with the LLM’s inherent ability to process and generate balanced and optimized code. Symmetry in this context allows for more efficient resource allocation and reduces redundant operations, further enhancing the model’s effectiveness. We believe that LLMs hold substantial potential in optimizing microservice deployment models.

Artificial intelligence models to predict optimal trade-off on construction management

PurposeThis research aims to introduce a novel algorithm, the Chaotic Giant Pacific Octopus Optimizer (CGPOO) and demonstrate framework includes four key aspects: time, cost, quality and safety trade-off (TCQST).Design/methodology/approachArtificial intelligence is causing a big disruption in the construction management. It is being used to building projects to enhance efficiency, safety and decision-making. This research compared the CGPOO method to those of other algorithms, such as the Chaotic Slime Mold Algorithm (CSMA), the Chaotic Salps Swarm Algorithm (CSSA) and the Chaotic Whale Optimization Algorithm (CWOA) and assessed the efficacy of the method using statistical analysis and evaluation indicators such as Hyper-volumn (HV), Spread (Sp), Computational Time (CT) and C-metric.FindingsThe analysis demonstrates that using CGPOO outperforms standalone methods chosen from the literature in terms of outcomes. It is discovered that the CGPOO solution possibilities for each factors are more efficient and beneficial than the comparison algorithms. Moreover, the CGPOO model performs better than the other algorithms with quality indices C-metric, Sp, HV and CT of 0.534, 0.531, 0.891 and 101.Originality/valueThe article presents a novel hybrid CGPOO that permits multi-factor trade-offs in construction management with the goal of surpassing the analyzed models and optimizing the optimal solution in the search space.

Self-correction of the optical distortion effect of thermal plumes in particle image velocimetry

Optical distortion caused by changes in the refractive index of fluid flow is a common issue in flow visualization using techniques, such as particle image velocimetry (PIV). In thermally driven convection, this distortion can severely interfere with PIV results due to the ubiquitous density and, therefore, refractive index heterogeneity in the fluid. The distortion also varies spatially and temporally, adding to the challenge. We propose a composite filter, the shadow-affected PIV region filter, which combines a series of conventional image filters to address this issue, focusing on optical distortion of thermal plumes in laminar flow. We verify the effectiveness of the filter using both synthetic particle images created from ray tracing and real particle images from the laboratory. For the first time, we effectively mitigate the optical distortion from plumes while preserving the in-plane plume velocity and overall flow pattern, with the PIV data alone. Our filter is efficient and does not require additional measurements, expensive ray tracing, or a large dataset to begin with. It can be extended to separate the flow field and the effect of optical distortion in other fluid experiments when the two components are visually distinct. Additionally, this filter can serve as a baseline algorithm for comparison when developing more advanced methods like neural networks.

Visual detection of drilling robot position for rockburst prevention in mining processing by a new image dehazing method

The positioning of drill pipe based on visual detection is a crucial aspect in achieving unmanned operation of drilling robot for rockburst prevention. However, the images directly collected from mechanized mining face are always polluted. In order to eliminate the interference of dust and haze in the image, an image dehazing method based on improved CycleGAN model is proposed in this paper. Firstly, a pipe image dehazing dataset for the rockburst prevention drilling robot is collected and established. Moreover, a generator architecture with a multi-scale U-shaped network structure is designed to improve the quality and accuracy of image recovery. A new reconstruction block is designed and an SK fusion layer is introduced to improve the feature extraction capability of the model, and a MU-CycleGAN network structure is constructed. Finally, an experimental platform for drill pipe image dehazing of the drilling robot for rockburst prevention is set up in the intelligent mining equipment laboratory. Experimental results showed that the PSNR and SSIM of the image dehazing model achieved 27.04 and 0.946, and the success rate of drill pipe grabbing has increased by 13.75%. Experimental results reveal that the proposed framework achieves superior image enhancement performance than the comparison algorithms.

Decomposition prediction fractional-order active disturbance rejection control deep Q network for generation control of integrated energy systems

The number of distributed renewable energy sources of integrated energy systems (IESs) are increasing. The volatility of renewable energy exacerbates active power imbalances and increase frequency deviations in power systems. This work proposes a decomposition prediction fractional-order active disturbance rejection control deep Q network (DPFOADRCDQN) to control each generator of IESs more accurately and to reduce systemic frequency fluctuation. According to the fluctuation size of frequency deviation, the sub-signal obtained after the prediction of the modal decomposition of frequency deviation is classified into big fluctuating signals and small fluctuating signals; fractional-order active disturbance rejection control (FOADRC) controls small fluctuating signals; deep Q network (DQN) controls big fluctuating signals. Frequency deviation signals in complex power systems with a large percentage of renewable energy have higher smoothness after modal decomposition, FOADRC can efficiently follow the small fluctuating signals, and DQN can learn the rich data features in the big fluctuating signals and give precise control commands. The DPFOADRCDQN, proportional-integral-derivative, Q-learning, fuzzy control, and the previous state-of-the-art reinforcement learning are simulated in IESs where the renewable energy generators output is 0 %, 20 %, 40 %, 65 %, and 80 % of all generators output, the results show the average frequency deviation, total generation cost, and carbon emission cost are at least 42.7 %, 0.92 %, and 0.93 % smaller than the comparison algorithm, respectively.

Deep predictive data representation model control for photovoltaic maximum power point tracking under partial shading conditions

As deep learning progresses in recent years, deep neural networks (DNNs) have penetrated applications in various industries. The global maximum power point (GMPP) tracking issue exists in photovoltaic (PV) systems. This study proposes a deep predictive data representation model control (PDRC)-based PV GMPP tracking method. The PDRC approach establishes a data representation control model centered on deep prediction by extracting the data features such as the current, voltage, and duty cycle of PV systems controllers on a time series basis. The proposed PDRC method can receive signals of voltage and current, then output duty cycle control signals. Thereby PDRC method can control PV output power to track GMPP efficiently. Finally, the proposed PDRC approach is evaluated with the perturbation observation method (P&O), the improved particle swarm algorithm (PSO-P&O), and the modified cuckoo algorithm (CS-P&O) to compare the PV GMPP tracking performance. In the simulation experiments, the PDRC approach increases the average tracking efficiency over the comparison algorithms by at least 7.729 % and decreases the average tracking time by at least 0.0155 s under partial shading conditions (PSCs). In the physical simulation validation, the proposed PDRC approach possesses minimal power perturbation and increases the average tracking efficiency over the comparison algorithms by at least 0.407 %.

Semi-Supervised Soft Computing for Ammonia Nitrogen Using a Self-Constructing Fuzzy Neural Network with an Active Learning Mechanism

Ammonia nitrogen (NH3-N) is a key water quality variable that is difficult to measure in the water treatment process. Data-driven soft computing is one of the effective approaches to address this issue. Since the detection cost of NH3-N is very expensive, a large number of NH3-N values are missing in the collected water quality dataset, that is, a large number of unlabeled data are obtained. To enhance the prediction accuracy of NH3-N, a semi-supervised soft computing method using a self-constructing fuzzy neural network with an active learning mechanism (SS-SCFNN-ALM) is proposed in this study. In the SS-SCFNN-ALM, firstly, to reduce the computational complexity of active learning, the kernel k-means clustering algorithm is utilized to cluster the labeled and unlabeled data, respectively. Then, the clusters with larger information values are selected from the unlabeled data using a distance metric criterion. Furthermore, to improve the quality of the selected samples, a Gaussian regression model is adopted to eliminate the redundant samples with large similarity from the selected clusters. Finally, the selected unlabeled samples are manually labeled, that is, the NH3-N values are added into the dataset. To realize the semi-supervised soft computing of the NH3-N concentration, the labeled dataset and the manually labeled samples are combined and sent to the developed SCFNN. The experimental results demonstrate that the test root mean square error (RMSE) and test accuracy of the proposed SS-SCFNN-ALM are 0.0638 and 86.31%, respectively, which are better than the SCFNN (without the active learning mechanism), MM, DFNN, SOFNN-HPS, and other comparison algorithms.

Research for an enhanced fault-tolerant solution against the current sensor fault types in induction motor drives

Introduction. Recently, three-phase induction motor drives have been widely used in industrial applications; however, the feedback signal failures of current sensors can seriously degrade the operation performance of the entire drive system. Therefore, the motor drives require a proper solution to prevent current sensor faults and improve the reliability of the motor drive systems. The novelty of the proposed research includes integrating the current sensor fault-tolerant control (FTC) function according to enhanced technique into the field-oriented control loop for speed control of the motor drive system. Purpose. This research proposes a hybrid method involving a third difference operator and signal comparison algorithm to diagnose various types of current sensor faults as a positive solution to enhance the stability of the induction motor drive system. Methods. A hybrid method involving a third difference operator for the measured speed signals and a comparison algorithm between measured and estimated current signals are proposed to diagnose the current sensors’ health status in the fault-tolerant process. After determining the faulty sensor, the estimated current signals based on the Luenberger observer are used immediately to replace the defective sensor signal. Results. The current sensor is simulated with various failure types, from standard to rare failures, to evaluate the performance of the FTC method implemented in the MATLAB/Simulink environment. Simultaneously, a fault flag corresponding to a defective sensor should be presented as an indicator to execute the repair process for faulty sensors at the proper time. Practical value. Positive results have proven the feasibility and effectiveness of the proposed FTC integrated into the speed controller to improve reliability and ensure the stable operation of the induction motor drive system even under current sensor fault conditions. References 29, tables 3, figures 10.

FLCMC: Federated Learning Approach for Chinese Medicinal Text Classification.

Addressing the privacy protection and data sharing issues in Chinese medical texts, this paper introduces a federated learning approach named FLCMC for Chinese medical text classification. The paper first discusses the data heterogeneity issue in federated language modeling. Then, it proposes two perturbed federated learning algorithms, FedPA and FedPAP, based on the self-attention mechanism. In these algorithms, the self-attention mechanism is incorporated within the model aggregation module, while a perturbation term, which measures the differences between the client and the server, is added to the local update module along with a customized PAdam optimizer. Secondly, to enable a fair comparison of algorithms' performance, existing federated algorithms are improved by integrating a customized Adam optimizer. Through experiments, this paper first conducts experimental analyses on hyperparameters, data heterogeneity, and validity on synthetic datasets, which proves that the proposed federated learning algorithm has significant advantages in classification performance and convergence stability when dealing with heterogeneous data. Then, the algorithm is applied to Chinese medical text datasets to verify its effectiveness on real datasets. The comparative analysis of algorithm performance and communication efficiency shows that the algorithm exhibits strong generalization ability on deep learning models for Chinese medical texts. As for the synthetic dataset, upon comparing with comparison algorithms FedAvg, FedProx, FedAtt, and their improved versions, the experimental results show that for data with general heterogeneity, both FedPA and FedPAP show significantly more accurate and stable convergence behavior. On the real Chinese medical dataset of doctor-patient conversations, IMCS-V2, with logistic regression and long short-term memory network as training models, the experiment results show that in comparison to the above three comparison algorithms and their improved versions, FedPA and FedPAP both possess the best accuracy performance and display significantly more stable and accurate convergence behavior, proving that the method in this paper has better classification effects for Chinese medical texts.

Adaptive group sample with central momentum contrast loss for unsupervised individual identification of cows in changeable conditions

Individual identification is the basis for accurate breeding and management of dairy cows. Most of the studies on individual identification of dairy cows focused on closed data sets. Meanwhile, the training process requires lots of manual labeling data as support. Aiming at solving the above shortcomings, an unsupervised cow individual identification model named AC-Cowid (Adaptive Group Sample-Central Momentum Contrastive loss-Cowid) was proposed. The innovation of the model lied in the novel sampler and loss function. Adaptive Group Sampler made up for the sampling imbalance problem in the current samplers. The sampler would self-update with model optimization. The central constraint term in Central Momentum Contrastive loss made the updated features more inclined to the cluster center, effectively solved the discomfort of Contrastive loss to cow image data. Feature similarity matching was adopted to meet the challenge of data conversion from a closed set to an open and changeable set. The unsupervised training reduced the workload of manual data labeling. The CMC-1 and mAP index of the final model were 99.50 % and 77.90 %, respectively. Ablation and hyper-parameter experiments proved the effectiveness of the proposed method. The comparison experiment tested eight unsupervised methods and five supervised algorithms in the state-of-the-art. The mAP and CMC-1 of the proposed method was the best among the comparison algorithms. Above all, the proposed method was 4.60 % and 5.38 % higher on mAP and CMC-1 compared with the best algorithm in supervised learning. This study could provide technical support for individual identification of cows in large-scale cattle farms.

Styled and characteristic Peking opera facial makeup synthesis with co-training and transfer conditional styleGAN2

Against the backdrop of the deep integration of culture and technology, research and practice in digitization of intangible cultural heritage has continued to deepen. However, due to the lack of data and training, it is still very difficult to apply artificial intelligence to the field of cultural heritage protection. This article integrates image generation technology into the digital protection of Peking opera facial makeup, using a self-built Peking opera facial makeup dataset. Based on the StyleGAN2 network, we propose a style generative cooperative training network Co-StyleGAN2, which integrates the adaptive data augmentation (ADA) to alleviate the problem of discriminator overfitting and introduces the idea of cooperative training to stabilize the training process. We design a Peking opera facial makeup image transform conditional generation network TC-StyleGAN2 which is transferred from unconditional generation network. The weights of the unconditional pre-training model are fixed, and an adaptive filtering modulation module is added to modulate the category parameters to complete the conversion from unconditional to conditional StyleGAN2 to deal with the training difficulty of conditional GANs on limited data, which suffer from severe mode collapse. The experimental results show that the proposed training strategy is better than the comparison algorithms, and the image generation quality and diversity have been improved.

Uncertain data density peak clustering algorithm based on JS divergence

Aiming at the defects of traditional density-based uncertainty clustering algorithms, such as parameter sensitivity and poor clustering results for complex manifold uncertain data sets, a new uncertainty data density peak clustering algorithm based on JS divergence (UDPC-JS) is proposed. The algorithm first removes noise points by using the uncertain natural neighbor density factor defined by the uncertain natural neighbor. Secondly, the local density of uncertain data objects is calculated by combining the uncertain natural neighbor and JS divergence. The initial clustering center of the uncertain data set is found by combining the idea of representative points, and the distance between the initial clustering centers is defined based on JS divergence and graph. Then, the decision graph is constructed on the initial clustering center using the local density calculated based on the uncertain natural neighbor and JS divergence and the newly defined distance between the initial clustering centers based on JS divergence and graph, and the final clustering center is selected according to the decision graph. Finally, the unassigned uncertain data objects are assigned to the cluster where their initial clustering center is located. Experimental results show that the algorithm has better clustering effect and accuracy than the comparison algorithm, and has a greater advantage in processing complex manifold uncertain data sets.

Isolation Forest Anomaly Detection Algorithm Based On Multi-level Sub-subspace Partition

In the research of credit loan fraud detection, the isolation forest algorithm has attracted much attention because of its ability to efficiently process large-scale data sets. However, when facing high-dimensional data, the performance of the isolation forest algorithm is easily affected, resulting in deviation of the detection results. In order to solve the above problems, this paper proposes an isolation forest anomaly detection algorithm based on multi-level sub-subspace division. Firstly, the random forest algorithm is used to evaluate the importance of each feature, and the data is divided into different subspaces according to the importance of each feature, and the corresponding weight is assigned to each subspace. Then, the isolation forest algorithm is applied in each subspace for anomaly detection, and the anomaly score of each subspace is obtained. Finally, the anomaly score and weight of each subspace were combined to obtain the final anomaly detection score. In order to evaluate the effectiveness of the algorithm, the proposed algorithm was compared with other four algorithms on the credit loan fraud data set. The results show that the AUC index, accuracy, recall rate and F1 score of the proposed algorithm are higher than those of the comparison algorithms, showing high effectiveness.

Adjudication of Codes for Identifying Sepsis in Hospital Administrative Data by Expert Consensus*

OBJECTIVES: Refine the administrative data definition of sepsis in hospitalized patients, including less severe cases. Design and Setting: For each of 1928 infection and 108 organ dysfunction codes used in Canadian hospital abstracts, experts reached consensus on the likelihood that it could relate to sepsis. We developed a new algorithm, called AlgorithmL, that requires at least one infection and one organ dysfunction code adjudicated as likely or very likely to be related to sepsis. AlgorithmL was compared with four previously described algorithms, regarding included codes, population-based incidence, and hospital mortality rates—separately for ICU and non-ICU cohorts in a large Canadian city. We also compared sepsis identification from these code-based algorithms with the Centers for Disease Control’s Adult Sepsis Event (ASE) definition. SUBJECTS: Among Calgary’s adult population of 1.033 million there were 61,632 eligible hospitalizations. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: AlgorithmL includes 720 infection codes and 50 organ dysfunction codes. Comparison algorithms varied from 42–941 infection codes to 2–36 organ codes. There was substantial nonoverlap of codes in AlgorithmL vs. the comparators. Annual sepsis incidence rates (per 100,000 population) based on AlgorithmL were 91 in the ICU and 291 in the non-ICU cohort. Incidences based on comparators ranged from 28–77 for ICU to 11–266 for non-ICU cohorts. Hospital sepsis mortality rates based on AlgorithmL were 24% in ICU and 17% in non-ICU cohorts; based on comparators, they ranged 27–38% in the ICU cohort and 18–47% for the non-ICU cohort. Of AlgorithmL-identified cases, 41% met the ASE criteria, compared with 42–82% for the comparator algorithms. CONCLUSIONS: Compared with other code-based algorithms, AlgorithmL includes more infection and organ dysfunction codes. AlgorithmL incidence rates are higher; hospital mortality rates are lower. AlgorithmL may more fully encompass the full range of sepsis severity.

Hybrid modeling with data enhanced driven learning algorithm for smart generation control in multi-area integrated energy systems with high proportion renewable energy

Multi-area integrated energy systems (MA-IES) with a high proportion of renewable energy have become a trend in social development. The access to a large number of distributed energy sources makes the safe, stable, and economic operation of MA-IES suffer serious challenges. To reduce the frequency deviation and area control error (ACE) of MA-IES and to improve the economy of operation, this study proposed a hybrid modeling with data enhanced driven learning algorithm, which combines model-driven and data-driven algorithms, named combined proportion-integral-derivative and deep reinforcement learning (CPIDDRL). Firstly, the frequency deviation signals are collected and judged, and the smaller frequency deviation signals, which are sent to the model-driven part, are regulated by proportion-integral-derivative (PID) controllers. Then, the ACE and large frequency deviation signals are collected and transmitted to the data-driven part, and the Transformer deep learning network is applied to predict the multi-feature timeseries, which are applied to strategy selection for state-action-reward-state-action reinforcement learning. Finally, the model-driven part and data-driven part work together to generate adjustment commands every 4 s. The control effects of CPIDDRL, PID, Q-learning, and sliding model control algorithms are compared in two and four areas integrated energy systems. The results show that, compared to the comparison algorithms, the CPIDDRL reduces the mean values of frequency deviation and ACEs by at least 46.78% and 6.83%, respectively, and the total generation cost by at least 8.20%. CPIDDRL has practical significance in improving system stability, enhancing renewable energy integration capabilities, and promoting the development of smart grids.

A framework of integrated differential evolution variants based on adaptive relay mode for global optimization

Differential evolution (DE) is highly competitive in single-objective real parameter optimization. However, there are still some problems with differential evolutionary variants in optimizing complex multimode functions, such as poor solution accuracy and slow convergence. To address these problems, an optimization framework of integrated DE variants based on adaptive relay mode (fDE-ARM) is proposed. In this framework, different DE variants are integrated through two adaptive relay mechanisms to give full play to the advantages of the algorithms, thereby improving the performance of the whole algorithm. For the first adaptive relay mode, when the currently executed algorithm is judged to have converged, the population is updated by Gaussian random walk, and then the optimization is continued through the relay algorithm. For the second adaptive relay mode, the relay condition is determined by the average fitness improvement rate. If the condition is met, the relay algorithm takes over the optimization. At the same time, if the diversity is lower than the threshold, to improve the exploration, the roulette wheel method is used to select some individuals to perform the Gaussian random walk. In addition, a feedback mechanism is introduced in the relay process to avoid false switching. To verify the performance of the proposed algorithm, extensive simulations are performed on CEC2005, CEC2014, CEC2017, and CEC2021 benchmark functions. In addition, three engineering problems are used to test the performance. Compared with some newly proposed optimization algorithms, fDE-ARM is statistically superior to the comparison algorithms in terms of solution accuracy, convergence speed, and stability.