Swarm and Evolutionary Computation最新文献

{"title":"Reinforcement learning-integrated evolutionary algorithm for enhanced unmanned aerial vehicle coverage path planning","authors":"Seung Chan Choi ,&nbsp;Yohan Lee ,&nbsp;Sung Won Cho","doi":"10.1016/j.swevo.2025.102051","DOIUrl":"10.1016/j.swevo.2025.102051","url":null,"abstract":"<div><div>The rapid development of unmanned aerial vehicle (UAV) technologies has led to their increased utilization across various industries. In search and rescue (SAR) missions, UAVs play a critical role in overcoming mobility constraints in search environments, particularly in time-sensitive situations such as maritime operations. To enhance the efficiency of search missions, this study addresses the Coverage Path Planning (CPP) problem for multiple UAVs in irregularly shaped search areas. We propose a novel CPP framework consisting of two main phases. In Phase 1, a reinforcement learning-integrated evolutionary algorithm is introduced for search area decomposition, aiming to minimize the area of the grid map exceeding the search area. Specifically, proximal policy optimization-based particle swarm optimization (PPO–PSO) is employed to effectively adapt to complex and irregular shapes. In Phase 2, a Mixed Integer Linear Programming (MILP) model is formulated to minimize mission completion time while ensuring collision avoidance and efficient task allocation for multiple UAVs. The proposed methodology was validated through 15 experimental scenarios, including real-world maritime environments, and demonstrated superior performance compared to existing methods in managing irregularly shaped search areas.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"97 ","pages":"Article 102051"},"PeriodicalIF":8.2,"publicationDate":"2025-07-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144579286","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":"Optimum settings for discrete PID control of nonlinear systems","authors":"Robert Vrabel","doi":"10.1016/j.swevo.2025.102052","DOIUrl":"10.1016/j.swevo.2025.102052","url":null,"abstract":"<div><div>This study investigates the application of piecewise affine approximation techniques for the control of nonlinear systems, focusing on the effective linearization of systems described by the <span><math><mi>k</mi></math></span>th order difference equation <span><math><mrow><mi>x</mi><mrow><mo>[</mo><mi>n</mi><mo>+</mo><mi>k</mi><mo>]</mo></mrow><mo>+</mo><mi>f</mi><mrow><mo>(</mo><mi>x</mi><mrow><mo>[</mo><mi>n</mi><mo>]</mo></mrow><mo>,</mo><mi>x</mi><mrow><mo>[</mo><mi>n</mi><mo>+</mo><mn>1</mn><mo>]</mo></mrow><mo>,</mo><mo>…</mo><mo>,</mo><mi>x</mi><mrow><mo>[</mo><mi>n</mi><mo>+</mo><mi>k</mi><mo>−</mo><mn>1</mn><mo>]</mo></mrow><mo>)</mo></mrow><mo>=</mo><mi>u</mi><mrow><mo>[</mo><mi>n</mi><mo>]</mo></mrow></mrow></math></span>. The proposed approach employs piecewise linearization by partitioning the nonlinear function <span><math><mi>f</mi></math></span> into simplices within a compact domain <span><math><mrow><mi>D</mi><mo>⊂</mo><msup><mrow><mi>R</mi></mrow><mrow><mi>k</mi></mrow></msup></mrow></math></span>. The parameter <span><math><mi>h</mi></math></span>, which determines the number of linear segments, governs the precision of the approximation. As <span><math><mi>h</mi></math></span> increases, the linearized system’s behavior converges uniformly to that of the original nonlinear system, facilitating improved control system performance.</div><div>A key advantage of the approach is that it does not require full knowledge of the nonlinear function <span><math><mi>f</mi></math></span>; only values at selected nodal points are needed. Furthermore, it is sufficient that <span><math><mi>f</mi></math></span> is twice continuously differentiable within each subdomain of the partition. If bounds on the gradient and Hessian of <span><math><mi>f</mi></math></span> are available within each cell, the total approximation error can be rigorously estimated.</div><div>In addition, the study incorporates PID controllers and leverages the Particle Swarm Optimization (PSO) algorithm to optimize controller parameters. The optimization framework is designed to minimize key performance indices, such as the Integral Time Absolute Error (ITAE) and Integral Squared Overshoot (ISO). Numerical simulations demonstrate the efficacy of the proposed method, highlighting its ability to balance computational complexity with approximation accuracy in nonlinear control system design.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"97 ","pages":"Article 102052"},"PeriodicalIF":8.2,"publicationDate":"2025-07-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144571722","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":"The difficulty of predicting behavior on stochastic local search algorithms","authors":"Daniel Loscos,&nbsp;Narciso Martí-Oliet,&nbsp;Ismael Rodríguez","doi":"10.1016/j.swevo.2025.102010","DOIUrl":"10.1016/j.swevo.2025.102010","url":null,"abstract":"<div><div>Identifying key properties of Stochastic Local Search (SLS) algorithms, such as convergence to optimal solutions, is essential. Unfortunately, due to their Turing-completeness and Rice’s theorem, their non-trivial semantic properties are generally undecidable. Therefore, most convergence results are achieved by abusing properties that ultimately depict them as simple (probabilistic) exhaustive search algorithms. We show that the general difficulty to prove properties of SLS algorithms has a strong theoretical basis: even when SLS algorithms are deterministic and their memory is linearly bounded, finding out their output from their input configuration is PSPACE-hard — and thus intractable if P<span><math><mo>≠</mo></math></span>PSPACE. This is proven by translating the PSPACE-hard DLBA-ACCEPT problem (i.e. given a Deterministic Linear Bounded Automaton and a word, checking whether the automaton accepts the word) into an instance of the tile-matching problem MPCP such that its solution denotes the configurations traversed by the DLBA during its execution. Simple SLS algorithms can obtain increasing partial solutions for these MPCP instances and provide the answer of the original DLBA-ACCEPT instances. It is also shown that finding out whether an SLS algorithm using linear memory fulfills any non-trivial semantic property is PSPACE-hard. An adaptation of Rice’s theorem dealing with computation artifacts running with linear space is introduced for that purpose. In order to provide an intuitive test of PSPACE-hardness for SLS algorithms, examples of how our criteria is applied to several heuristics, such as depth-first search and genetic algorithms, are shown.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"97 ","pages":"Article 102010"},"PeriodicalIF":8.2,"publicationDate":"2025-07-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144571717","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":"Broad reinforcement learning based adaptive state transition algorithm for global optimization","authors":"Yangyi Du,&nbsp;Xiaojun Zhou,&nbsp;Chunhua Yang,&nbsp;Weihua Gui","doi":"10.1016/j.swevo.2025.102038","DOIUrl":"10.1016/j.swevo.2025.102038","url":null,"abstract":"<div><div>The state transition algorithm (STA) is an efficient intelligent optimization method with superior search capabilities in diverse applications, while its key operator selection strategies depend on manual design. The integration of deep reinforcement learning (DRL) with STA offers a promising paradigm for adaptive selection strategy during optimization. However, conventional DRL methods require extensive training data and iterative model refinement, creating fundamental barriers with limited evaluation budgets. Therefore, this paper proposes a novel STA framework incorporating broad reinforcement learning to develop an adaptive operator selection mechanism. First, the selection strategy is formulated as a Markov decision process, where an agent learns to identify optimal operators based on real-time state. Specifically, environmental states are characterized through systematic landscape analysis derived from population information. Second, a broad learning system replaces neural networks in DRL frameworks. The associated incremental learning mechanism is carefully designed to enhance training efficiency. Third, a Gaussian mixture model-based data augmentation mechanism is proposed to generate sufficient training samples under limited interactions. The proposed method is evaluated using benchmark functions and practical applications, with comparisons against STA variants and other prominent optimization algorithms. Experimental results demonstrate that BRL-STA achieves competitive performance compared with competitors.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"97 ","pages":"Article 102038"},"PeriodicalIF":8.2,"publicationDate":"2025-07-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144571724","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":"Population decomposition evolutionary framework for constrained multiobjective optimization","authors":"Yongchao Li ,&nbsp;Heming Jia ,&nbsp;Hongguang Li","doi":"10.1016/j.swevo.2025.102055","DOIUrl":"10.1016/j.swevo.2025.102055","url":null,"abstract":"<div><div>The solution to constrained multiobjective optimization problems (CMOPs) requires both optimizing the objective function and satisfying the constraints. Many studies have demonstrated that multi-population models are effective for solving CMOPs. However, excessive consumption of evaluation times can lead to convergence difficulties in the later stages of population evolution.This article proposes a population decomposition strategy to overcome these drawbacks and enhance the quality of the solution set. Specifically, clustering techniques partition both the main and unconstrained populations in the objective space, yielding <span><math><mi>r</mi></math></span> subpopulations and, consequently, <span><math><mrow><mi>r</mi><mo>+</mo><mn>1</mn></mrow></math></span> subpopulations. A fuzzy selection mechanism is introduced to enhance offspring convergence while preserving population diversity. By reformulating the selection of the optimal individual as a conditional extremum problem within a fuzzy environment, the algorithm’s applicability to CMOPs is significantly improved. Additionally, a novel environmental selection model for unconstrained populations is proposed to ensure both convergence and diversity. In the early stage, this model prioritizes convergence by leveraging the Euclidean distance in the target space. In the later stage, diversity is maintained by incorporating both Euclidean distance and cosine similarity. Finally, comparisons with six state-of-the-art constrained multiobjective evolutionary algorithms on 57 benchmark test functions and 12 real-world problems demonstrate that the proposed algorithm achieves superior performance in terms of both convergence and diversity. The code for PDECMO is <span><span>https://github.com/YongchaoLucky/PDECMO.git</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"97 ","pages":"Article 102055"},"PeriodicalIF":8.2,"publicationDate":"2025-07-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144571723","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 dynamic multi-objective evolutionary greedy algorithm for distributed hybrid flow shop rescheduling problem","authors":"Xin-Rui Tao ,&nbsp;Quan-Ke Pan ,&nbsp;Xue-Lei Jing ,&nbsp;Wei-Min Li","doi":"10.1016/j.swevo.2025.102054","DOIUrl":"10.1016/j.swevo.2025.102054","url":null,"abstract":"<div><div>This paper addresses a distributed hybrid flowshop rescheduling problem (DHFRP) with new job insertion and machine breakdowns, which exists widely in modern industry. The objective is to minimize makespan and total tardiness time. A dynamic multi-objective evolutionary greedy algorithm is used to solve this rescheduling problem. An initialization strategy is designed to generate a high-quality initial population. An adaptive perturbation process and a local search procedure further enhance the quality of the population. For dynamic changes in the processing environment, the valid information in the original non-dominated solution set is utilized in order to efficiently obtain a rescheduling solution. In addition, a deep reinforcement learning algorithm is used to make decisions on the rescheduling strategy to be adopted. This operation maintains the stability of the production process and effectively reduces the transportation cost during processing. Finally, a series of numerical experiments demonstrate the effectiveness of the proposed algorithm in solving DHFRP. This is further supported by a case study based on real industrial production.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"97 ","pages":"Article 102054"},"PeriodicalIF":8.2,"publicationDate":"2025-07-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144557654","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 decimal artificial bee colony with elite strategy for the cutting stock problem with irregular items","authors":"Qiang Luo ,&nbsp;Chunrong Pan ,&nbsp;Hong Zhong ,&nbsp;Yunqing Rao","doi":"10.1016/j.swevo.2025.102026","DOIUrl":"10.1016/j.swevo.2025.102026","url":null,"abstract":"<div><div>This study investigates an irregular cutting stock problem in various industrial applications, including shipbuilding, construction machinery, and automobiles, where a considerable quantity of metal sheets are consumed. The problem involves cutting the single-size stocks to produce a set of demanded items such that the material utilization is maximized, i.e., the waste is minimized. To address the problem, this study employs the double scanline to represent the irregular items, and proposes a decimal artificial bee colony with elite strategy. The algorithm represents solutions with decimal vectors and uses a decoder procedure to map these vectors to solutions of the problem. In addition, a metaheuristic-based hybrid algorithm is developed for further improving the solution quality. To comprehensively assess the performance of the algorithm, two sets of computational tests were conducted. The experimental results demonstrated that the proposed algorithm outperforms competing algorithms by achieving faster convergence than other metaheuristics of the same class and producing better solutions, verifying the algorithm's effectiveness and superiority. The implementation of the algorithm benefits waste reduction for companies in practice.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"97 ","pages":"Article 102026"},"PeriodicalIF":8.2,"publicationDate":"2025-07-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144534427","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 adaptive hybrid neighborhood search algorithm for the electric vehicle pickup and delivery problem with time windows and partial charging","authors":"Saiqi Zhou ,&nbsp;Dezhi Zhang ,&nbsp;Shiyan Fang ,&nbsp;Shuangyan Li","doi":"10.1016/j.swevo.2025.102042","DOIUrl":"10.1016/j.swevo.2025.102042","url":null,"abstract":"<div><div>Environmental pressures and measures are compelling the extensive integration of electric vehicles into transportation and logistics systems. This paper focuses on addressing the electric vehicle pickup and delivery problem with time windows and partial charging, in which the amount of charging electricity at charging stations is flexible and determined based on the route schedules. A new effective mixed-integer linear programming model has been developed for the problem. To effectively tackle large-scale instances, we propose an adaptive hybrid neighborhood search algorithm, which is based on the framework of the adaptive large neighborhood search algorithm. The proposed algorithm incorporates various problem-oriented search operators being adaptively chosen for evolution. Meanwhile, dynamic programming-based charging approaches for both full and partial charging policies are presented. Numerical experiments are conducted using benchmark instances of the electric vehicle pickup and delivery problem to verify the effectiveness of our algorithm configurations and its overall performance. The solution results are compared against those obtained using the state-of-the-art algorithm, and the proposed algorithm identifies 21 new best solutions and exhibits greater stability, which demonstrates the competitiveness of the proposed algorithm. Furthermore, the analysis of charging policies provides interesting insights, highlighting the significant advantage of the partial charging policy in scenarios characterized by clustered customer distributions or short scheduling horizons.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"97 ","pages":"Article 102042"},"PeriodicalIF":8.2,"publicationDate":"2025-07-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144534428","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":"Decomposition-based multi-objective reinforcement learning for dynamic disassembly job shop scheduling with urgency guidance","authors":"Fangyu Li,&nbsp;Ruichong Ma,&nbsp;Jiarong Du,&nbsp;Honggui Han","doi":"10.1016/j.swevo.2025.102040","DOIUrl":"10.1016/j.swevo.2025.102040","url":null,"abstract":"<div><div>The dynamic disassembly job shop scheduling problem (DDJSSP) entails organizing multiple jobs with distinct requirements across machines, where job operations are subject to sequence constraints and urgency conditions. Existing deep reinforcement learning techniques for multi-objective job shop scheduling problems (JSSP) with manually weighted reward functions result in a single policy, limiting the ability to approximate multiple policies in the Pareto front. To minimize both makespan and total energy consumption in DDJSSP, we propose an urgency-driven decomposition-based multi-objective reinforcement learning (UD-MORL) approach. We simulate a dynamic scheduling environment reflecting real-world complexities by introducing uncertain processing times, random employee absences, and an urgency rate for orders. We then develop a decomposition approach to separate objectives by adjusting weights and iterating policies based on performance and information entropy metrics. Finally, we employ a mutual information mechanism to identify the weight combination exhibiting the strongest correlation with population points, thereby improving weight-fitting efficiency. Experimental results on public general-purpose JSSP datasets show UD-MORL outperforms existing multi-objective reinforcement learning algorithms in hypervolume and sparsity, achieving an average hypervolume improvement, sparsity reduction, and a win-rate of 55% across all benchmark instances.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"97 ","pages":"Article 102040"},"PeriodicalIF":8.2,"publicationDate":"2025-07-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144522297","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":"Solving nonlinear equation systems and real-world engineering example via adaptive information migration and sharing evolutionary multitasking algorithm with cross-sampling","authors":"Zhihui Fu ,&nbsp;Suruo Li","doi":"10.1016/j.swevo.2025.102059","DOIUrl":"10.1016/j.swevo.2025.102059","url":null,"abstract":"<div><div>In practical engineering problems, nonlinear equation systems (NESs) are widely present, such as power systems and control systems. With the increase of system scale and the complexity of problems, solving these NESs becomes increasingly difficult. Although existing methods have proposed effective improvement methods from multiple perspectives, they still ignore the key issue that the implicit relationship between different NESs can promote the evolution of algorithms. Therefore, this paper proposes adaptive cross-sampling evolutionary multitasking algorithm framework, namely AC-MTNESs, to solve NESs. This framework establishes the implicit relationship between tasks through unified encoding method, and proposes an adaptive information migration and sharing selection mechanism, combined with fractional calculus methods, to more accurately capture the nonlinear relationship between NESs. In addition, to ensure that the algorithm maintains the level of population diversity, this work propose a cross-resource sampling strategy, which balances the exploration and exploitation capabilities of the algorithm by archiving roots that meet the accuracy threshold and reusing resources from different distributions to cross-generate offspring. Experiments verify the superiority of the algorithm on 30 standard and 18 complex NESs problem sets. The results show that AC-MTNESs outperforms existing methods. Furthermore, it also shows good application potential in practical problems of motor systems.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"97 ","pages":"Article 102059"},"PeriodicalIF":8.2,"publicationDate":"2025-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144522296","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}