Abstract Satellite imagery plays a crucial role in exploring land use inventories of urban areas. However, accurate land cover classification from satellite imagery remains a longstanding challenge. With recent advancements in artificial intelligence technology, Deep Learning algorithms have achieved success in understanding satellite images by means of Convolutional Neural Networks (CNNs). While there has been a notable emphasis on satellite image analysis to improve the accuracy of land cover classifications, it is imperative to emphasise the significance of data-driven optimisation techniques. This paper introduces a hybrid UNet-ResNet-50 architecture, which integrates the metaheuristic Particle Swarm Algorithm (PSA) in dynamic hyperparameter optimisation for multi-class semantic segmentation. The approach of this research leverages a UNet extractor with ResNet-50 backbone (UResNet-50) and augments it with a Particle Swarm Optimiser (PSO) to automate the hyperparameter tuning process for segmenting the DeepGlobe satellite dataset into seven meaningful classes, namely: urban, forest, rangeland, barren land, agriculture, water bodies and unknown. The PSO-UResNet-50 model demonstrated robust performance across four distinct locations, in terms of accuracy, precision, recall, F1-score and mIoU as follows: Location-1 (95.74%, 98.12%, 86.95%, 92.04%,88.17%); Location-2 (91.88%, 79.23%, 80.75%, 81.42%, 83.03%); Location-3 (99.44%, 93.97%, 87.42%, 88.68%, 90.77%); and Location-4 (96.20%, 94.03%, 89.75%, 92.16%, 88.97%). The proposed PSO-UResNet-50 model outperformed the conventional U-Net and hybrid UResNet-50, demonstrating the advantage of applying PSO in multi-class segmentation of satellite imagery. The principal contribution of this work lies in the development and validation of a novel, metaheuristic-optimised deep learning framework that addresses the land cover classification challenge inherent in satellite images.

Head-mounted gradient coils for brain imaging need to be designed on irregular surfaces. However, current gradient coil design methodologies often struggle to produce optimal magnetic field solutions for these specialized geometries. The stream function method is typically not applicable to irregular two-dimensional surfaces, while the boundary element method involves excessive variables that are difficult to optimize. To develop a new gradient coil design method with fewer variables suitable for irregular surfaces, and to complete the design of gradient coils for brain imaging on a semi-ellipsoidalsurface. Triangular meshes are constructed on the target surface, which is then harmonically mapped to a rectangle. After obtaining the mapping relationship, the connection between the stream function on the rectangle and the coil shape on the original surface is established via the inverse of the harmonic map. Finally, the particle swarm algorithm is used to optimize the basis function coefficients and complete the coil design. The method was used to design and coils on a semi-ellipsoid surface. The resulting coils were manufactured and measured to verify the effectiveness of the method. Both simulation and measurement results show that the and coils designed with this method can generate more linear magnetic fields with roughly the same inductance. The maximum deviation of the magnetic field generated by the Gy coil is 4.02% (simulated) and 15.78% (measured), while that by the compared method is 33.3%. The maximum deviation of the magnetic field generated by the Gz coil is 2.57% (simulated) and 7.20% (measured), while that by the compared method is 27.8%. This paper provides a new method for coil design on irregular two-dimensional single connected bounded surface, to design gradient coil with better magnetic field with only a few basis functions. The linearity and uniformity of the obtained magnetic field have significant advantages over the boundary element method. Even with manufacturing and measurement errors amplifying the deviation, the measured deviation remains far lower than that of the referencedesign.

Radial piston hydraulic motors are widely used in engineering attributed to their lightweight, compactness, and high efficiency. However, internal vibration and shock during operation can influence the performance of the motor. In this study, the stator curve composition and motion characteristics of radial piston hydraulic motors were investigated. Specifically, an improved particle swarm algorithm was extracted to optimize the internal curve of radial piston hydraulic motors with adaptive nonlinear dynamic inertia weights, a good point-set initial population strategy, and a competitive optimization strategy with multiple sub-populations. Additionally, iso-acceleration curves with compensation zones were reconstructed to minimize abrupt changes in shock and contact stresses. The optimized stator was subjected to force analysis with the method of maximum stress-life analysis. Furthermore, the motor fatigue life was accurately evaluated by finite element analysis. The results suggest that the enhanced particle swarm algorithm, which achieved a 34.63% increase in optimization efficiency, can be utilized to design radial piston hydraulic motor stators swiftly and effectively. Moreover, the optimization of the stator brought about a 27% increase in the motor's actual lifespan.

ABSTRACT This study presents an integrated parametric optimization framework combining the Taguchi method with a multi-objective particle swarm optimization (MOPSO) algorithm to enhance the welding performance of an industrial robotic gas metal arc welding (GMAW) system. The Taguchi method is first employed to identify the most influential parameters and their preliminary optimal levels for welding quality, based on orthogonal array experimentation. These results are then used as the initial population for the MOPSO algorithm, which performs a refined global search to obtain Pareto-optimal solutions balancing multiple, often conflicting, objectives. The optimization focuses on three critical quality indices: depth-to-width (D/W) ratio, heat input, and wire consumption. Experimental results using S400 low-carbon steel demonstrate that the proposed hybrid method outperforms the conventional Taguchi-only approach, achieving a 6.17% increase in D/W ratio, a 17.63% reduction in heat input, and a 15.68% decrease in wire consumption. The findings confirm that the Taguchi-MOPSO hybrid approach effectively improves weld quality, reduces energy and material usage, and enhances overall process efficiency in robotic welding applications.

This paper provides a brief introduction to building integrated photovoltaic (BIPV). It utilized a particle swarm algorithm to optimize the arrangement scheme of the photovoltaic array on a concrete flat roof. Furthermore, the performance of the particle swarm algorithm was improved through a genetic algorithm (GA). Subsequently, an experimental analysis was conducted using an office building in a joint salinization factory in Henan Province. First, the effectiveness of the building energy model constructed using PKPM-Energy and Daysim software was validated. Then, a comparison was made between the GA, the particle swarm algorithm, and the improved particle swarm algorithm. The results indicated that the building energy consumption model constructed by the simulation software calculated the building’s energy consumption effectively, which was used to guide the optimization algorithm. All three optimization algorithms successfully optimized the initial scheme. Moreover, the improved particle swarm algorithm exhibited the best optimization performance among the three algorithms.

This study focuses on the process conditions for the preparation of $C4$ olefins via ethanol-catalyzed coupling technology for the wide range of applications of $C4$ olefins in the production of chemicals and pharmaceuticals, a process of remarkable significance and value for the optimization of chemical products. In order to maximize the yield of $C4$ olefin under the given experimental conditions, this paper combines the BP neural network model of Bayesian optimization with the black-box optimization model, and further solves for the optimal catalyst combinations and reaction temperatures by the particle swarm algorithm with the introduction of a variational strategy. Through this comprehensive method, the key conditions for achieving the maximum yield of $C4$ olefins were determined, demonstrating the good application of this method in the optimization process of the chemical industry.

Gear ultrasonic vibration grinding composite function curve distribution heat source model optimized based on particle swarm algorithm

Reactive power optimization of power systems based on an improved particle swarm algorithm

The creation of a powerful intrusion detection system (IDS) is important due to the security implications of 6G wireless communications and ultra-densified networks that exceed the capabilities of existing systems. IDSs are inadequate to prevent ongoing unidentified attacks on wireless networks. In this manuscript, anomaly detection in communication systems via self-competition swarm optimized hypergraph neural networks with a self-competition particle swarm optimization algorithm (CSHNN-SCPSOA-DA-CS) is proposed. At first, the input data are collected from the CIC-IDS2017 dataset. The input data are pre-processed using Bayesian boundary trend filtering (BBTF) to remove unwanted data. The pre-processed data are then fed to the weighted leader search algorithm (WLSA), which is employed to select optimal features. The selected features are then provided to anomaly detection using central-smoothing hypergraph neural networks (CSHNN) for classifying and detecting attacks such as cross-site scripting (XSS), brute force, SQL injection, port scan, web attacks, botnet, infiltration, and distributed denial-of-service (DDoS). Hence, the self-competition particle swarm optimization algorithm (SCPSOA) is used to optimize the CSHNN, which accurately classifies anomaly detection. The proposed CSHNN-SCPSOA-DA-CS approach achieves 19.51%, 22.95%, and 30.56% higher accuracy; 16.60% compared to existing techniques such as SVM-AD-CN, FL-AN-IDS, and CNN-DAA-NBAD, respectively.

With the increasing penetration of renewable energy, power systems are facing greater uncertainty and volatility, which poses significant challenges for Virtual Power Plant scheduling. Existing research mainly focuses on optimizing economic efficiency but often overlooks system reliability and the impact of forecasting deviations on scheduling, leading to suboptimal performance. Thus, this paper presents a reliability-cost bi-objective cooperative optimization model based on a dual-swarm particle swarm algorithm: it introduces positive and negative imbalance price penalty factors to explicitly describe the economic costs of forecast deviations, constructs a reliability evaluation system covering PV, EVs, air-conditioning loads, electrolytic aluminum loads, and energy storage, and solves the multi-objective model via algorithm design of “sub-swarms specializing in single objectives + periodic information exchange”. Simulation results show that the method ensures stable intraday operation of VPPs, achieving 6.8% total cost reduction, 12.5% system reliability improvement, and 14.8% power deviation reduction, verifying its practical value and application prospects.

Meshing characteristics analysis for cycloidal mechanism of rotate vector reducer and optimized design based on particle swarm algorithm

Single-phase-to-ground faults occur frequently in distribution networks, while traditional localization methods have limitations such as insufficient feature extraction and poor topological adaptability. To address these issues, this paper proposes a two-stage localization method that integrates the Node Classification Matrix (NCM) and an Improved Binary Particle Swarm Optimization (IBPSO) algorithm. The NCM achieves rapid initial localization, and the IBPSO performs error correction. This paper employs an IEEE 33-node standard distribution network model to design simulations covering scenarios with varying fault locations, multiple fault resistances, and different numbers of node distortions for validation. The results demonstrate that the proposed method achieves a fault location accuracy of 96%, which is 19% higher than that of the NCM alone and 2% higher than that of the IBPSO alone. Moreover, it maintains an accuracy of over 95% under scenarios of 1–3 node distortions, topological switching, and high-impedance faults, and is compatible with existing Feeder Terminal Unit (FTU) devices. This method effectively balances localization speed and robustness, providing a reliable solution for the rapid fault isolation of distribution network.

The development of 18 GHz hybrid superconducting ECR ion sources is constrained by the complex trade-off in magnet system design, where achieving simultaneous excellence in field strength, confinement stability, and resonant coupling remains a formidable challenge. A design automation framework that tightly integrates Particle Swarm Optimization (PSO) with COMSOL-based finite element analysis is presented. This synergy enables the global optimization of the permanent magnet hexapole and the superconducting solenoids’ currents as an interconnected system. The optimizer delivers a magnetic field configuration that simultaneously achieves a 2.6 T axial peak, a 4.25 mirror ratio, and a precise minimum-B field of 0.6 T. This synergy creates a stable magnetic cage perfectly resonant at 18 GHz, ensuring superior plasma confinement and efficient microwave-to-plasma energy transfer. This study validates the PSO algorithm as a powerful tool for transcending conventional design paradigms in complex electromagnetic systems. The resulting magnet solution not only meets the stringent demands of next-generation ECR ion sources but also provides a transferable blueprint for optimizing a broad class of symmetric devices governed by multi-physics constraints.

Optimizing Star-Delta Starter Transitions for Induction Motors with Particle Swarm Algorithm

On‐ramp merging is a common expressway maneuver for connected and automated vehicles (CAVs), where trajectory planning and tracking control are central to avoiding collisions. However, existing studies rarely optimize the selection of merge start and end points and give limited attention to constraints from acceleration‐lane length. This study proposes a structured trajectory planning and tracking method with multiobjective optimization under the CAV’s environment. First, by sampling the starting and ending points of the merging process, the quintic polynomial is used to plan the initial trajectory of the merging vehicles, and trajectory safety is checked with a collision‐avoidance algorithm based on rectangular vehicle geometry. Then, a multiobjective optimization model selects the on‐ramp trajectory by balancing merging urgency, driving safety, traffic efficiency, and comfort. Finally, an integrated tracking strategy combines lateral and longitudinal control: a feedforward LQR for lateral motion and a PID‐based longitudinal controller. To further improve the tracking accuracy, the particle swarm algorithm tunes key parameters of the lateral LQR controller. The numerical result demonstrates that the planner can generate smooth and stable trajectories that could be selected as an optimal reference for the tracking controller. The simulation results show that when the initial speed of the on‐ramp vehicle is 68 km/h, the maximum tracking errors of lateral and longitudinal displacements are less than 0.02 and 0.2 m, respectively.

A hybrid proximal policy optimization and particle swarm algorithm for highway alignment optimization

In low-voltage networks, phase-unbalanced loads are caused by the presence of single-phase consumers. Furthermore, due to the presence of railway traction substations powered by two phases in the power supply system, there is a problem of electromagnetic compatibility between the traction power supply system and the rest of the electrical power system. An unbalanced mode of operation results in additional energy losses, deterioration in power supply quality, reduced system efficiency and stability, and decreased service life of devices. For most low-voltage networks, the zero-sequence voltage asymmetry coefficient typically exceeds the requirements of power quality standards. To optimize the network operating mode with asymmetric loads, this paper proposes to apply load balancing by phases. The objective of this study is a comparative analysis of algorithms for balancing loads across phases in low-voltage networks to select the most effective one according to the criterion of minimum total costs while observing the specified restrictions on the asymmetry coefficient and voltage levels in the nodes. Materials and methods. An optimization method based on a heuristic algorithm – the particle swarm algorithm – was applied. Power flows and voltage values ​​at the nodes were calculated on a per-phase basis, taking into account active power and voltage losses. The particle swarm algorithm was implemented in the Python programming language. Results. To balance loads, two objective functions have been considered. The first function included the total network operating costs, voltage deviation at the nodes, and the zero-sequence voltage asymmetry coefficient. The second function comprised the total network operating costs and voltage deviation at the nodes. A study of two optimization algorithms has been conducted: based on sensitivity coefficients and the particle swarm algorithm; as well as solely on the particle swarm algorithm. The first algorithm comprised two stages: identifying potential nodes for balancing with the help of sensitivity coefficients, and then determining device power using the particle swarm algorithm. The second algorithm determined the device installation nodes and their power applying the particle swarm algorithm. The results obtained from the two optimization algorithms and two different objective functions have been analyzed. The values ​​of the asymmetry coefficients of zero-sequence voltage and voltages in network nodes in pre-optimization and post-optimization modes have been investigated. According to the obtained results, the lowest total costs will be achieved with two-stage optimization (with the help of sensitivity coefficients and the particle swarm algorithm). This optimization requires an objective function that includes the total cost index and the voltage deviation index at the nodes. Conclusions. When optimizing a network for load balancing, the algorithm applying sensitivity coefficients in combination with the particle swarm algorithm proves to be effective. The objective function should include total network operating costs and voltage deviations. This optimization ensures reduction of active power losses by 12.8% of the pre-optimization losses and decrease of total network operating costs.

Abstract The integration of renewable energy into modern power systems requires scheduling strategies that balance economic efficiency with low-carbon objectives. This study develops an optimal scheduling framework for virtual power plant (VPP) operating under carbon trading and green certificate mechanisms. First, the model coordinates the dispatch of gas turbines, wind power, photovoltaic units, and energy storage systems, incorporating market incentives and emission constraints into a unified optimization problem. Second, a Q-learning enhanced particle swarm optimization algorithm (QPSO) is designed, which adaptively adjusts inertia weights and learning factors according to search states to improve convergence stability and solution quality. Finally, comparative analyses with the standard particle swarm optimization algorithm and independent Q-learning demonstrate significant improvements: under the dual-market scenario, net profit increases by 89.9%, renewable energy utilization rises by 19.9%, and carbon emissions are reduced by 39.4%. These results indicate that combining dual-market participation with adaptive optimization provides a feasible and effective approach to enhancing both the economic and environmental performance of VPP operations.

In multi-objective particle swarm optimization (MOPSO), challenges persist, including low diversity in external archives, ambiguous individual optimal choice mechanisms, high sensitivity to parameter settings, and the arduous task of balancing global exploration and local exploitation capabilities. To address these issues, this paper introduces a novel multi-objective particle swarm optimization algorithm named HCRMOPSO. The proposed algorithm innovatively leverages hierarchical clustering based on Ward’s linkage to generate the center of mass as reference points, which are then combined with the ideal point and crowding distance. This effectively maintains the external archive, thereby resolving the diversity deficiency commonly found in traditional MOPSO archives. Additionally, HCRMOPSO fuses multiple particles to update the personal best positions. It also adaptively tunes the flight parameters according to the diversity information within each particle’s neighborhood, enhancing the algorithm’s adaptability. Notably, a new strategy is designed for two specific types of particles, further optimizing the search process. The performance of HCRMOPSO is rigorously evaluated against ten existing algorithms on 22 standard test problems. Experimental results demonstrate that HCRMOPSO outperforms its counterparts on multiple benchmarks, showcasing superior effectiveness in handling multi-objective optimization tasks.

Optimization of centrifugal fan for oolong tea shaking machine based on Kriging model and multi-objective particle swarm algorithm.