Swarm and Evolutionary Computation最新文献

{"title":"Surrogate-assisted evolutionary algorithm with stage-adaptive infill sampling criterion for expensive multimodal multi-objective optimization","authors":"Yufei Yang ,&nbsp;Changsheng Zhang ,&nbsp;Yi Liu ,&nbsp;Haitong Zhao","doi":"10.1016/j.swevo.2025.102068","DOIUrl":"10.1016/j.swevo.2025.102068","url":null,"abstract":"<div><div>The key issue in handling expensive multimodal multi-objective optimization problems is to balance convergence and diversity in both the decision and objective spaces with limited function evaluations available. To tackle this issue, this paper proposes a surrogate-assisted multimodal multi-objective evolutionary algorithm with stage-adaptive infill sampling criterion. In the proposed algorithm, a multi-surrogate cooperative framework is developed, where multiple extreme gradient boosting models are used to approximate the objective functions for replacing real function evaluations, and a self-organizing map (SOM) network is used to learn the topologies of Pareto sets in the decision space and corresponding features in the objective space for reducing the approximation errors. Then, a stage-adaptive infill sampling criterion is designed to select the most suitable candidates for expensive function evaluations. Specifically, in the first stage, a convergence-first infill sampling criterion is used to accelerate convergence to the global Pareto front; In the second stage, an indicator-based infill sampling criterion according to neuron weights of the SOM network and a diversity-based infill sampling criterion are used to improve diversity in decision and objective spaces. Experimental results on two benchmark test suites demonstrate the competitiveness of the proposed algorithm against eight state-of-the-art methods.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"98 ","pages":"Article 102068"},"PeriodicalIF":8.2,"publicationDate":"2025-07-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144654319","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Multi-objective energy-efficient scheduling of distributed heterogeneous hybrid flow shops via multi-agent double deep Q-Network","authors":"Minghai Yuan,&nbsp;Yang Ye,&nbsp;Hanyu Huang,&nbsp;Zhen Zhang,&nbsp;Fengque Pei,&nbsp;Wenbin Gu","doi":"10.1016/j.swevo.2025.102076","DOIUrl":"10.1016/j.swevo.2025.102076","url":null,"abstract":"<div><div>The Distributed Heterogeneous Hybrid Flow Shop Scheduling Problem (DHHFSP) poses a highly complex NP-hard combinatorial optimization challenge, particularly under the dual pressures of green manufacturing and distributed production. To address the limitations of conventional optimization algorithms in handling large-scale multi-objective scheduling with dynamic constraints, this paper proposes a novel Multi-Agent Double Deep Q-Network (MADDQN) based energy-efficient scheduling framework. The DHHFSP is formulated as a Markov Decision Process (MDP), where hierarchical agents representing jobs, workshops, and machines collaboratively learn optimal scheduling policies through a shared state representation and discrete rule-based action spaces. A hybrid reward mechanism combining delayed immediate rewards and global rewards is designed to efficiently guide the agents toward minimizing due time error (DTE) and total energy consumption (TEC). In addition, an adaptive energy-saving strategy is introduced to further reduce standby energy consumption without compromising delivery deadlines. Extensive computational experiments demonstrate that the proposed MADDQN achieves superior performance over state-of-the-art algorithms such as NSGA-II and optimal single rule methods in terms of convergence, solution diversity, and computational efficiency, with average improvements of 59.76%, 67.25%, and 99.72%, respectively. Furthermore, Pareto-based multi-objective evaluation metrics are utilized to comprehensively assess the balance between conflicting objectives. An industrial case study validates the practical applicability of the proposed method within real-world manufacturing execution systems (MES), offering a scalable and intelligent solution for energy-efficient scheduling in distributed heterogeneous manufacturing environments.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"98 ","pages":"Article 102076"},"PeriodicalIF":8.2,"publicationDate":"2025-07-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144654318","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"When robots learn from nature: GLWOA-RRT*, a nature-inspired motion planning approach","authors":"Sara Bouraine ,&nbsp;Yassine Bellalia ,&nbsp;Ilyes Chaabeni ,&nbsp;Djamila Naceur","doi":"10.1016/j.swevo.2025.102062","DOIUrl":"10.1016/j.swevo.2025.102062","url":null,"abstract":"<div><div>Optimal motion planning for autonomous mobile robots remains still an open issue, with the classical approaches often struggle to address real-world conditions imposed on the system. In recent years, there has been a surge of interest in nature-inspired approaches in solving different technological problems by imitating natural processes. In this context, the present paper concerns the development of a motion planner based on a nature-inspired approach, which is dubbed GLWOA-RRT*. It is based on the Rapidly Exploring Random Tree Star (RRT*), offering an efficient exploration of the search space for a good initialisation of agents, and the Whale Optimisation Algorithm (WOA), one of the most efficient nature-based approaches thanks to its fast convergence, high calculation accuracy, efficiency in balancing exploration and exploitation to avoid falling into a local optimum, and demonstrated remarkable performance in tackling real-world challenges in different fields. GLWOA-RRT* is conceptually built as a global approach, solving the motion planning problem by encoding each agent in the swarm by the robot’s motion. The problem is solved by determining the optimal and safe motion to be executed by the robot. GLWOA-RRT* has been applied first in simulation in both static and dynamic environments, where a detailed experimental analysis of WOA parameters in a motion planning context has been performed. Furthermore, it has been tested in a range of challenging scenarios, where the results demonstrate the efficiency of the algorithm in finding a valid and optimal motion in a reasonable time. The obtained results also exhibit the performance of GLWOA-RRT* in improving RRT* by offering better results, especially in challenging scenarios. Finally, GLWOA-RRT* has been implemented and validated in real-world scenarios using the Robot Operating System (ROS), both in 3D Gazebo-based simulation environments and in physical environments. The outcomes demonstrate its efficacy and applicability in real-world settings.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"98 ","pages":"Article 102062"},"PeriodicalIF":8.2,"publicationDate":"2025-07-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144632266","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A finite-element dual-level contextual informed neural network with swarm space hopping algorithm based optimal feature selection and detection for EEG-based epileptic seizure detection","authors":"Nagarjuna Tandra ,&nbsp;Nikhat Akhtar ,&nbsp;K Padmanaban ,&nbsp;L. Guganathan","doi":"10.1016/j.swevo.2025.102072","DOIUrl":"10.1016/j.swevo.2025.102072","url":null,"abstract":"<div><div>This study proposes a novel approach for Electroencephalogram (EEG) based Epileptic Seizure Detection (ESD) using a Finite-Element Dual-Level Contextual Informed Neural Network (AFi-EDLCINNet) integrated with the Swarm Space Hopping Algorithm (SSHA). The approach addresses the challenges of contextual sensitivity and computational efficiency in current ESD methods. Raw EEG signals from the Children's Hospital Boston-Massachusetts Institute of Technology (CHB-MIT) and Bonn datasets are preprocessed with Weighted Guided Image Filtering and Entropy Evaluation Weighting (WGIF-EEW) to eliminate noise and ensure signal clarity. Feature extraction is performed using Multi-Discrete Wavelet Transform (MDWT) to capture critical patterns. A hybrid optimization method combining the White Shark Optimizer and Brown Bear Optimization Algorithm (WSO-BBOA) is used for optimal feature selection, making certain that just the most important features are included are selected. The selected features are input into AFi-EDLCINNet for classification, which is further optimized by SSHA to improve accuracy and efficiency in detecting epileptic seizures. The proposed method achieves an impressive 99.9 % classification accuracy and a low error rate of 0.7 %, outperforming other methods. This framework offers a reliable, robust solution for early seizure detection, providing clinicians with a powerful tool for personalized treatment planning. The solution is implemented using Python, demonstrating its practicality and flexibility for real-world applications.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"97 ","pages":"Article 102072"},"PeriodicalIF":8.2,"publicationDate":"2025-07-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144623481","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A large-scale multi-objective algorithm integrating prisoner’s dilemma model and prospect theory with adaptive learning","authors":"Yu Sun ,&nbsp;Wenhao Cai","doi":"10.1016/j.swevo.2025.102061","DOIUrl":"10.1016/j.swevo.2025.102061","url":null,"abstract":"<div><div>Numerous multi-objective problems contain many decision variables in the real world, which are referred to as large-scale multi-objective problems (LSMOPs). The challenge of achieving a balance between convergence and diversity in large-scale multi-objective optimization is addressed in this paper, which introduces a hybrid strategy for solving LSMOPs. Initially, a two-stage adaptive learning strategy based on history is employed to update the population. Subsequently, the prisoner’s dilemma model in game theory is introduced during the offspring generation stage to balance convergence and diversity. A fuzzy evolutionary mechanism is then employed for optimization to enhance the diversity and searchability of the population. Finally, a reference vector-guided selection and a risk preference mechanism based on prospect theory are employed to perform selection during the environmental selection phase. Experimental results on benchmark problems with 100–1000 decision variables reveal that the algorithm has the best overall performance compared with state-of-the-art large-scale multi-objective evolutionary algorithms (MOEAs).</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"97 ","pages":"Article 102061"},"PeriodicalIF":8.2,"publicationDate":"2025-07-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144632458","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Bi-level Voronoi strategy for cooperative search and coverage","authors":"Benyamin Ebrahimi ,&nbsp;Ali Asghar Bataleblu ,&nbsp;Jafar Roshanian","doi":"10.1016/j.swevo.2025.102064","DOIUrl":"10.1016/j.swevo.2025.102064","url":null,"abstract":"<div><div>In this paper, a bi-level Voronoi-based path planning strategy is proposed to address the challenge of cooperative multi-agent search and coverage in uncertain environments. While traditional Voronoi-based coverage control is commonly utilized for optimal path planning, its limitations, such as agents' premature convergence to Voronoi centroids, leading to reduced exploration and lack of incentive to move, can hinder system efficiency. The proposed bi-level strategy provides a framework to overcome such limitations while ensuring a more balanced and adaptive allocation of the environment among agents, thereby enhancing overall performance in terms of environmental mean uncertainty reduction and target detection. This framework utilizes a primary Voronoi diagram based on agent positions for initial spatial partitioning. To enhance exploration efficiency, a secondary Voronoi tessellation is applied, integrating probabilistic information about the target’s existence. The bi-level framework enables agents to achieve purposeful coverage by employing an efficient Voronoi partition allocation that integrates both the agents' positions and the probability of target existence. To this end, a novel allocation approach is employed to assign Voronoi neighbors to agents, ensuring that common cells within each agent's region are allocated to the most deserved agent. This mechanism promotes proportional contributions to uncertainty reduction, ensuring that each agent prioritizes areas of higher uncertainty or greater target likelihood. By doing so, agents operate efficiently, effectively reducing environmental uncertainty and improving target detection. Simulation results and comparative analyses validate the proposed strategy, demonstrating its superiority over conventional methods and highlighting its significance in multi-agent cooperative missions.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"97 ","pages":"Article 102064"},"PeriodicalIF":8.2,"publicationDate":"2025-07-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144613940","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A reinforcement learning-enhanced multi-objective Co-evolutionary algorithm for distributed group scheduling with preventive maintenance","authors":"Yanan Wang ,&nbsp;Yuyan Han ,&nbsp;Yuting Wang ,&nbsp;Hongyan Sang ,&nbsp;Yuhang Wang","doi":"10.1016/j.swevo.2025.102066","DOIUrl":"10.1016/j.swevo.2025.102066","url":null,"abstract":"<div><div>In the context of the global impetus towards sustainable development and in response to grow market demands, there is a critical need for multi-regional, multi-objective, and flexible production models. Under this background, this article first formulates a mathematical model of a distributed group scheduling problem with preventive maintenance (DFGSP_PM), in which the machine's maintenance level drops below a preset threshold, preventive maintenance is triggered. Second, a reinforcement learning-enhanced multi-objective co-evolutionary algorithm (QCMOEA) is proposed. It incorporates a collaborative evaluation mechanism tailored to the characteristics of the coupled problems to extensively explore the solution space. To retain a balance between convergence and distribution properties, a solution selection strategy based on double-rank and cosine similarity approaches is utilized. Additionally, a Q-learning mechanism is adopted to dynamically select the optimal strategy during enhancing evolution for the group population. Furthermore, a three-stage increasing efficiency and reducing consumption strategy is designed by dynamically changing the machine speed. Finally, by conducting a comparative analysis with four existing metaheuristic algorithms across 405 test cases, the proposed algorithm demonstrates superior optimization capabilities in addressing this complex DFGSP_PM problem.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"97 ","pages":"Article 102066"},"PeriodicalIF":8.2,"publicationDate":"2025-07-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144597412","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"An optimized multi-source feature extraction model with black-winged kite algorithm for hourly seamless PM2.5 estimation","authors":"Li Wang ,&nbsp;Lili Xu ,&nbsp;Shurui Fan ,&nbsp;Yong Zhang","doi":"10.1016/j.swevo.2025.102069","DOIUrl":"10.1016/j.swevo.2025.102069","url":null,"abstract":"<div><div>Fine particulate matter (PM_2.5) estimation is becoming of significant importance for public health protection and air quality management. However, it is difficult to extract effective data features and easily lead to data missing with single-source dataset. And the multi-source data PM_2.5 estimation method will also increase data dimensionality and nonlinear data coupling. In this paper, a novel hourly seamless PM_2.5 estimation method is proposed with leveraging multi-source data integration through attentive interpretable tabular learning network (TabNet)-based feature extraction and the categorical boosting (CatBoost) prediction model. Firstly, a multi-source dataset is constructed for the satellite and reanalysis data fusion to solve the missing spatio-temporal data problem and to realize the hourly seamless estimation. Secondly, the TabNet structural parameters are optimized by the black-winged kite algorithm (BKA), effectively extracting key feature information from the multi-source data. And then the hourly seamless PM_2.5 concentration is estimated by CatBoost with Bayesian optimization (BO), which improves the generalization performance of the whole model. Finally, the multi-source dataset of Beijing-Tianjin-Hebei (BTH) region is used for the experimental analysis and validation. The results demonstrate that BKA has superiority over other algorithms in optimizing TabNet feature extraction. The proposed method exhibits excellent performance in hourly seamless PM_2.5 estimation, with the R² of the day and night models reaching 0.91 and 0.92 and the RMSE as low as 12.22 μg/m³ and 11.70 μg/m³, which are better than other regression models. In addition, the hourly PM_2.5 estimated are generally consistent with observations and the applicability across various spatial locations is validated.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"97 ","pages":"Article 102069"},"PeriodicalIF":8.2,"publicationDate":"2025-07-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144597411","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Performance optimization of Band Pass Filters and Wavelength Division Multiplexers for communication systems","authors":"Arup Kumar Ghosh ,&nbsp;Subhankar Bhattacharjee ,&nbsp;Gautam Garai","doi":"10.1016/j.swevo.2025.102049","DOIUrl":"10.1016/j.swevo.2025.102049","url":null,"abstract":"<div><div>The growing demand for compact, high-speed, and spectrally precise components in next-generation communication systems poses significant challenges in the design and optimization of photonic Band Pass Filters (BPFs) and Wavelength Division Multiplexers (WDMs). Conventional algorithms, such as Genetic and Taguchi Optimization (GTO), Particle Swarm Optimization (PSO), Grasshopper Optimization (GRO), and Bald Eagle Search Optimization (BESO), often suffer from premature convergence, spectral inaccuracies, limited flexibility, and poor scalability when applied to complex, high-dimensional photonic structures. To address these limitations, this study introduces the Advanced Distributed Dynamic Differential Evolution (AD³E) algorithm, an advanced distributed optimization technique. Applied to 1D Si_yGe_1-y–SiO₂ photonic crystals, AD³E achieves outstanding results in BPFs with 99. 9814% Transmittivity and a 0.4 nm FWHM at the center wavelength of 1550 nm and WDMs with 99. 9579% Transmittivity, a 0.4 nm FWHM and 0.6 nm channel spacing to avoid crosstalk. The dynamic parameter adaptation of AD³E significantly outperforms static settings. Further validation shows less than 8% performance degradation under <span><math><mo>±</mo></math></span>5% fabrication tolerance while Wilcoxon signed-rank testing (<span><math><mrow><mi>p</mi><mo>&lt;</mo><mn>0</mn><mo>.</mo><mn>05</mn></mrow></math></span>) confirms its statistical superiority over GTO, PSO, GRO, and BESO in 32 benchmark functions (30D and 100D). AD³E stands out as a powerful tool for next-generation photonic device optimization.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"97 ","pages":"Article 102049"},"PeriodicalIF":8.2,"publicationDate":"2025-07-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144579285","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"An improved discrete multi-objective artificial protozoa optimizer for solving multi-objective knapsack problems","authors":"Junjie Liao ,&nbsp;Zheng-Ming Gao ,&nbsp;Syam Melethil Sethumadhavan ,&nbsp;Gaoshuai Su ,&nbsp;Juan Zhao","doi":"10.1016/j.swevo.2025.102070","DOIUrl":"10.1016/j.swevo.2025.102070","url":null,"abstract":"<div><div>The multi-objective knapsack problem (MOKP) is a challenging combinatorial optimization problem that traditional methods often fail to solve effectively. Consequently, researchers are increasingly adopting metaheuristic algorithms to address such problems within a reasonable time. This paper introduces an improved discrete multi-objective artificial protozoa optimizer (IDMOAPO) to tackle MOKP. The continuous solution space of the leaded sine cosine multi-objective artificial protozoa optimizer is discretized using two approaches, among which the modulo operation is identified as the most effective and adopted to develop a discrete multi-objective artificial protozoa optimizer (DMOAPO). An enhanced strategy is further incorporated into DMOAPO to improve solution quality, resulting in the development of the proposed IDMOAPO. The proposed IDMOAPO is evaluated across 16 MOKPs of four types and compared against seven algorithms. The performance metrics used for the evaluation are the number of Pareto solutions, generational distance, Spread, and inverted generational distance. Simulation results show that IDMOAPO significantly outperforms other comparison algorithms in most cases. These results highlight the effectiveness of IDMOAPO in obtaining superior Pareto fronts, confirming its suitability for solving MOKP.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"97 ","pages":"Article 102070"},"PeriodicalIF":8.2,"publicationDate":"2025-07-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144589064","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}