Adaptive multi-objective swarm intelligence for containerized microservice deployment

A cooperative hybrid breeding swarm intelligence algorithm for feature selection

Multi-objective swarm intelligence approach for bias mitigation in decision-making software

Design of construction project management technology based on project schedule cost model and swarm intelligence algorithm

Automated guided vehicle (AGV) path planning aims to obtain an optimal path from the start point to the target point. Path planning methods are generally divided into classical algorithms and reactive algorithms, and this paper focuses on reactive algorithms. Reactive algorithms are classified into swarm intelligence algorithms and artificial intelligence algorithms, and this paper reviews relevant studies from the past six years (2019–2025). This review involves 123 papers: 81 papers are about reactive algorithms, 44 are based on the swarm intelligence algorithm, and 37 are based on artificial intelligence algorithms. The main categories of swarm intelligence algorithms include particle swarm optimization, ant colony optimization, and genetic algorithms. Neural networks, reinforcement learning, and fuzzy logic represent the main trends in artificial intelligence–based algorithms. Among the cited papers, 45.68% achieve online implementations, and 33.33% address multi-AGV systems. Swarm intelligence algorithms are suitable for static or simplified dynamic environments with a low computational complexity and fast convergence, as 79.55% of papers are based on a static environment and 22.73% achieve online path planning. Artificial intelligence algorithms are effective for dealing with dynamic environments, which contribute 72.97% to online implementation and 54.05% to dynamic environments, while they face the challenge of robustness and the sim-to-real problem.

WSNs have become significant in contemporary applications in the field of environmental monitoring, health care, and industrial automation. But these are limited by the fact that they are not powerful enough. In order to achieve a longer life span of networks and to ensure that data is delivered with greater reliability, there is a need to design routing protocols that are less energy-consuming. The paper suggests a new swarm intelligence routing algorithm, which combines both adaptive clustering and energy-efficient route optimization to optimize intra- and inter-cluster communication. The proposed approach is founded on the way swarm agents work in concert and modify routing decisions by residual energy, communication cost, and node density to maintain the energy consumption of the network within reasonable bounds. The simulation findings prove that our method enhances both the network lifetime and the ratio of packet delivery, achieves a throughput better than the traditional protocols, and minimizes the quantity of control overhead. This single-method design offers a resource-limited and rigorous scheme of resource-limited WSNs, which makes it especially suitable for realistic applications in energy-sensitive scenarios.

This article presents a description of a software complex for intelligent decision support in managing welding work organization and implementing a technology to improve the efficiency of these processes using the developed software complex. The complex integrates modules for welders’ training, task assignment, quality control, and automatic detection of welding seam defects. The core element of intelligent decision support is implementing adaptive swarm intelligence algorithms (bee colony, ant colony, firefly algorithms) for personalized training trajectories and staff qualification improvement, as well as evolutionary modelling for rational distribution of welding assignments among workers. Integration of defect data detected by convolutional neuro-fuzzy networks ensures a closedloop feedback cycle, increasing the objectivity of work quality assessments and the validity of managerial decisions. Applying the technology for improving the efficiency of managing welding work organization using the developed software complex comprises six consecutive stages, starting from defect identification and ending with task redistribution adjustments after training. The developed approach improves the quality of welds, reduces labour intensity, and shortens execution timelines.

With the explosive growth of data across various fields, effective data preprocessing has become increasingly critical. Evolutionary and swarm intelligence algorithms have shown considerable potential in feature selection. However, their performance often deteriorates in large-scale problems, due to premature convergence and limited exploration ability. To address these limitations, this paper proposes an algorithm named IHBOFS, a biomimetics-inspired optimization framework that integrates multiple adaptive strategies to enhance performance and stability. The introduction of the Good Point Set and Elite Opposition-Based Learning mechanisms provides the population with a well-distributed and diverse initialization. Furthermore, adaptive exploitation–exploration balancing strategies are designed for each subpopulation, effectively mitigating premature convergence. Extensive ablation studies on the CEC2022 benchmark functions verify the effectiveness of these strategies. Considering the discrete nature of feature selection, IHBOFS is further extended with continuous-to-discrete mapping functions and applied to six real-world datasets. Comparative experiments against nine metaheuristic-based methods, including Harris Hawk Optimization (HHO) and Ant Colony Optimization (ACO), demonstrate that IHBOFS achieves an average classification accuracy of 92.57%, confirming its superiority and robustness in high-dimensional feature selection tasks.

The Ivy Algorithm (IVYA), a swarm intelligence algorithm inspired by plant growth, presents a novel framework for optimization. To unlock its full potential in complex, high-dimensional problems, it is crucial to address the fundamental challenge of balancing exploration and exploitation, which can impact overall search efficiency and solution quality. To this end, this paper proposes an Enhanced Ivy Algorithm (E-IVYA) that integrates three synergistic mechanisms. First, a dynamic perturbation framework combining symmetric and asymmetric exploration is introduced to maintain population diversity. Second, a dynamic escape mechanism based on elite differential mutation is employed to prevent search stagnation and effectively escape from local optima. Third, an adaptive movement strategy inspired by the Sine-Cosine Algorithm is integrated to achieve a more adaptive balance between global exploration and local exploitation. The performance of the proposed E-IVYA was rigorously evaluated through two distinct phases. Initially, its optimization capabilities were benchmarked against a wide range of classic and advanced algorithms on the challenging IEEE CEC 2014 and 2017 test suites. Subsequently, its practical utility was validated by applying it to the complex task of automating the hyperparameter optimization of Generative Adversarial Networks (GANs) for imbalanced data classification. The experimental results demonstrate E-IVYA's superior performance. On the standard benchmarks, E-IVYA consistently ranked as a top-performing algorithm. In the practical application, the E-IVYA-optimized GAN model achieved a minority class F1-Score of 0.87 on the highly imbalanced Credit-Card Fraud dataset, significantly outperforming models augmented with standard techniques like SMOTE (0.71). These findings confirm that E-IVYA is a robust and efficient tool for tackling complex optimization problems, particularly in the domain of automated machine learning.

Gamification Education (GE), by combining game concepts like game elements with scaffolding and other teaching methods, can effectively stimulate students’ learning interest, cognitive thinking, and emotional expression. However, due to the diversity of student groups, individuals within and between groups also exhibit unique characteristics due to differences in cognitive abilities and other factors. These differences lead to conflicting educational goals, hindering the effectiveness of GE in education. Therefore, this paper uses a literature review method to study the generation, manifestation, and application of the contradiction between group differences and individual characteristics in GE. After that, it proposes a framework based on Universal Design for Learning (UDL) and a dynamic process based on adaptive learning, combining the two through swarm intelligence. UDL increases the breadth of gamified instructional design, ensuring it fully accounts for the differences and needs of groups, while adaptive instruction increases the depth of general education, tapping into the unique characteristics of each individual within the group. Through the collective discussion and screening of swarm intelligence, combined with UDL and adaptive instruction, it can achieve a balance between group differences and individual characteristics.

A key limitation of existing swarm intelligence (SI) algorithms for Node Coverage Optimization (NCO) is their inadequate solution accuracy. A novel chaotic quantum-inspired leader honey badger algorithm (CQLHBA) is proposed in this study. To enhance the performance of the basic HBA and better solve the numerical optimization and NCO problem, an adjustment strategy for parameter to balance the optimization process of the follower position is used to improve the exploration ability. Moreover, the chaotic dynamic strategy, quantum rotation strategy, and Lévy flight strategy are employed to enhance the overall performance of the designed CQLHBA, especially for the exploitation ability of individuals. The performance of the proposed CQLHBA is verified using twenty-one benchmark functions and compared to that of other state-of-the-art (SOTA) SI algorithms, including the Honey Badger Algorithm (HBA), Chaotic Sea-Horse Optimizer (CSHO), Sine–Cosine Quantum Salp Swarm Algorithm (SCQSSA), Golden Jackal Optimization (GJO), Aquila Optimizer (AO), Butterfly Optimization Algorithm (BOA), Salp Swarm Algorithm (SSA), Grey Wolf Optimizer (GWO), and Randomised Particle Swarm Optimizer (RPSO). The experimental results demonstrate that the proposed CQLHBA exhibits superior performance, characterized by enhanced global search capability and robust stability. This advantage is further validated through its application to the NCO problem in wireless sensor networks (WSNs), where it achieves commendable outcomes in terms of both coverage rate and network connectivity, confirming its practical efficacy in real-world deployment scenarios.

5G private power networks require high levels of real-time performance and reliability; however, traditional threshold-based scaling methods struggle with the unpredictable bursts and periodic fluctuations in traffic. This paper proposes a BiLSTM-Attention model that incorporates Swarm Intelligence-based Multi-Agent Collaborative Prediction (SI-MACP), named BiLSTM-Attention-SI-MACP, for resource forecasting in these networks. The proposed framework combines bidirectional LSTM networks with attention mechanisms to capture multivariate resource dependencies. Meanwhile, the SI-MACP mechanism utilizes principles of swarm intelligence through distributed task decomposition, privacy-preserving local modeling, and the collaborative aggregation of parameters and predictions. Experimental results demonstrate significant improvements, including 32% MAE reduction, 35% higher resource utilization, and 99.9% SLA compliance, particularly benefiting latency-sensitive services like differential protection.

ABSTRACT Hydrogen energy has many potential applications in manufacturing, transportation, and other industries. However, demand forecasts for industrial hydrogen are limited and unreliable. Based on the source of production, hydrogen energy can be divided into three categories: “gray hydrogen,” “blue hydrogen,” and “green hydrogen.” Among these, “gray hydrogen” is produced using fossil fuels and has high carbon emissions; “blue hydrogen” combines carbon capture and storage technology and has relatively low carbon emissions; “green hydrogen” is produced using renewable energy electrolysis and has zero carbon emissions. The study in this paper focuses on green hydrogen, which is hydrogen produced by electrolysis of water. This paper investigates the demand for industrial hydrogen using ammonia production as a case study, and it creates an intelligent forecast model for hydrogen demand. This model is built on PSO, which enhances RF, and an Optimized Least Squares Support Vector Machine modified with the Sparrow Algorithm. Firstly, this paper uses text mining to create a library of influencing factor indicators. The essential components are filtered out using a Random Forest improved on the particle swarm technique to prioritize their importance. Secondly, the paper presents an intelligent projection of the demand for industrial hydrogen in ammonia synthesis, with forecast findings ranging from 2024 to 2035. This enables us to calculate the total hydrogen demand in the industrial sector. Finally, an empirical analysis is performed using industrial hydrogen ammonia synthesis data from the national database. The intelligent prediction model proposed in this paper achieves the lowest MAPE of 7.77% and RMSE of 401.06 tons, superior to other comparison models. By 2035, ammonia demand is projected to reach 56.5 million tons, requiring 1527.23 tons of industrial hydrogen. The results show that the method described in this paper is more accurate and appropriate for estimating industrial hydrogen demand.

The autonomous navigation of UAV swarms in complex environments remains a critical bottleneck for their real-world, cost-effective deployment. This paper introduces a novel metaheuristic swarm intelligence framework, instantiated in the Self-Learning Slime Mould Algorithm (SLSMA), to address this challenge. The SLSMA empowers a UAV fleet to autonomously sense, adapt, and recover from planning failures through three core innovations: a situation-aware search strategy, a collective memory mechanism, and an adaptive recovery behavior. Rigorous evaluation on the CEC 2017 benchmark suite demonstrates that SLSMA achieves the topmost rank in the Friedman test and delivers statistically significant improvements (Wilcoxon rank-sum test, p < 0.05) over eight state-of-the-art metaheuristics. In complex 3D path planning scenarios, SLSMA generates complete collision-free trajectories where competing algorithms fail, achieving a 99.5% mission success rate and reducing the total cost by up to 18.7% compared to the best-performing variant. This work holds substantial importance for the scientific community by providing a foundational shift from static optimization to adaptive, metaheuristic reasoning in swarm robotics. It establishes a new paradigm for resilient autonomy, directly addressing the critical challenge of algorithmic fragility in uncertain environments. The proposed framework offers a replicable architecture that could influence the design of future intelligent systems beyond UAV navigation. The results establish SLSMA not merely as an optimizer, but as a robust solver for resilient autonomy, paving the way for the next generation of autonomous swarms capable of persistent operation in cluttered and dynamic environments.

The volatility of agricultural commodity prices significantly affects market stability and financial market dynamics, especially during periods of economic uncertainty and global shocks. Accurate price prediction, however, remains challenging due to the complex, nonlinear characteristics of agricultural markets and the diverse range of influencing factors. To overcome these challenges, this study develops a novel price forecasting framework that combines advanced time series decomposition, swarm intelligence optimization, and deep learning techniques. The proposed framework employs successive variational mode decomposition (SVMD) to deconstruct the raw price data into multiple components, effectively capturing the underlying nonlinear patterns and dynamic features. These components are then fed into a CNN-augmented BiLSTM model, enhanced with an attention mechanism to extract both temporal dependencies and intricate data relationships. To fine-tune the model’s hyperparameters, this study introduces a multiple strategies dung beetle optimisation algorithm (MSDBO), which integrates four strategic modifications to improve the balance between global search, local exploration, and convergence efficiency. Using historical data from corn and wheat markets as case studies, the experimental findings demonstrate that the proposed SVMD-MSDBO-CNN-BiLSTM-A model significantly outperforms nine baseline approaches. Specifically, it reduces the Mean Absolute Percentage Error (MAPE) by 25.78% and 37.57%, respectively, and enhances directional accuracy (Dstat) by 1.15% and 14.53% compared to the top single models.

Accelerating neural architecture search for LSTM with a training-free method based on swarm intelligence and Neural Tangent Kernel

A robust logistic regression approach enhanced by hyperparameter optimization techniques through swarm intelligence and genetic algorithms: Advancing cancer diagnosis

A swarm intelligence framework in complex environments: Optimizing area coverage guidance and control

Machine learning based on a swarm intelligence algorithm and explainable AI for the prediction of reservoir temperature

Swarm intelligence optimization-based task assignment for multiple autonomous underwater vehicles