Swarm and Evolutionary Computation最新文献_第8页

Integrating asynchronous advantage actor–critic (A3C) and coalitional game theory algorithms for optimizing energy, carbon emissions, and reliability of scientific workflows in cloud data centers

IF 8.2 1区计算机科学

Swarm and Evolutionary Computation Pub Date : 2025-02-01 DOI: 10.1016/j.swevo.2024.101756

Mustafa Ibrahim Khaleel

{"title":"Integrating asynchronous advantage actor–critic (A3C) and coalitional game theory algorithms for optimizing energy, carbon emissions, and reliability of scientific workflows in cloud data centers","authors":"Mustafa Ibrahim Khaleel","doi":"10.1016/j.swevo.2024.101756","DOIUrl":"10.1016/j.swevo.2024.101756","url":null,"abstract":"<div><div>The growth of workflow as a service (WFaaS) has become more intricate with the increasing variety and number of workflow module applications and expanding computing resources. This complexity leads to higher energy consumption in data centers, negatively impacting the environment and extending processing times. Striking a balance between reducing energy and carbon emissions and maintaining scheduling reliability is challenging. While deep reinforcement learning (DRL) approaches have shown significant success in workflow scheduling, they require extensive training time and data due to application homogeneity and sparse rewards, and they do not always guarantee effective convergence. On the other hand, experts have developed various scheduling policies that perform well for different optimization goals, but these heuristic strategies lack adaptability to environmental changes and specific workflow optimization. To address these challenges, an enhanced asynchronous advantage actor–critic (A3C) method combined with merge-and-split-based coalitional game theory is proposed. This approach effectively guides DRL learning in large-scale dynamic scheduling issues using optimal policies from the expert pool. The merge-and-split-based method prioritizes computing nodes based on their preemptive characteristics and resource heterogeneity, ensuring reliability-aware workflow scheduling that maps applications to computing resources while considering the dynamic nature of energy costs and carbon footprints. Experiments on real and synthesized workflows show that the proposed algorithm can learn high-quality scheduling policies for various workflows and optimization objectives, achieving energy efficiency improvements of 7.65% to 19.32%, carbon emission reductions of 3.13% to 14.76%, and reliability enhancements of 17.22% to 41.65%.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"92 ","pages":"Article 101756"},"PeriodicalIF":8.2,"publicationDate":"2025-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143183596","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}

引用次数: 0

A multi-population competitive evolutionary algorithm based on genotype preference for multimodal multi-objective optimization

IF 8.2 1区计算机科学

Swarm and Evolutionary Computation Pub Date : 2025-02-01 DOI: 10.1016/j.swevo.2024.101826

Keyu Zhong , Fen Xiao , Xieping Gao

{"title":"A multi-population competitive evolutionary algorithm based on genotype preference for multimodal multi-objective optimization","authors":"Keyu Zhong , Fen Xiao , Xieping Gao","doi":"10.1016/j.swevo.2024.101826","DOIUrl":"10.1016/j.swevo.2024.101826","url":null,"abstract":"<div><div>Many existing multimodal multi-objective evolutionary algorithms (MMOEAs) exhibit poor performance in addressing multimodal multi-objective optimization problems (MMOPs), mainly due to limited genetic diversity in environmental selection. In this paper, we propose a multi-population competitive evolutionary algorithm based on genotype preference (MPCEA-GP) to solve MMOPs. Firstly, we propose a population selection strategy based on genotype preference to maintain the genetic diversity of the population. This strategy utilizes the spectral radius to assess the overall convergence quality of the population, rather than evaluating each individual separately, and favors selecting the population with the minimum spectral radius, thereby preserving the genotypes of both optimal and suboptimal individuals. Secondly, to address the challenge of diminished genetic diversity during the evolutionary process, we incorporate historical survival population with substantial genetic diversity into the competition between parent and offspring, and preferentially select individuals with significant genotype differences to recombine into a new population. By merging two selected populations, a joint population with sufficient genetic diversity is constructed. Finally, a genotype-phenotype-based fitness criterion is devised to evaluate the fitness of individuals. This criterion not only compares genotypes using the Pareto dominance principle but also concurrently considers both genotype and phenotype diversity, aiding the population in more precisely identifying individuals with both good convergence and diversity. Empirical results show that MPCEA-GP outperforms state-of-the-art MMOEAs for 40 chosen benchmark functions and two complex real-world applications.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"92 ","pages":"Article 101826"},"PeriodicalIF":8.2,"publicationDate":"2025-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143183607","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}

引用次数: 0

Evolutionary optimization via swarming dynamics on products of spheres and rotation groups

IF 8.2 1区计算机科学

Swarm and Evolutionary Computation Pub Date : 2025-02-01 DOI: 10.1016/j.swevo.2024.101799

Vladimir Jaćimović , Zinaid Kapić , Aladin Crnkić

引用次数: 0

Knee-oriented expensive many-objective optimization via aggregation-dominance: A multi-task perspective

IF 8.2 1区计算机科学

Swarm and Evolutionary Computation Pub Date : 2025-02-01 DOI: 10.1016/j.swevo.2024.101813

Junfeng Tang , Handing Wang , Yaochu Jin

{"title":"Knee-oriented expensive many-objective optimization via aggregation-dominance: A multi-task perspective","authors":"Junfeng Tang , Handing Wang , Yaochu Jin","doi":"10.1016/j.swevo.2024.101813","DOIUrl":"10.1016/j.swevo.2024.101813","url":null,"abstract":"<div><div>Given the costs to implement whole Pareto optimal solutions, users often prefer solutions of interest, like knee points, which represent naturally preferred solutions without a specific bias. Recent surrogate-assisted multi-objective evolutionary algorithms (SAMOEAs) incorporating knee identification techniques have been suggested, but most of them cannot find knee solutions for expensive many-objective optimization problems. This work proposes a Kriging-assisted evolutionary multi-task algorithm with aggregation-dominance. The aggregation-dominance approach identifies knee points on an estimated Pareto front, from which subproblems are created and solved in parallel via Kriging-assisted multi-task optimization for guiding search knee solutions. Additionally, our proposed infill solutions selection strategy focuses on re-evaluating solutions converging in regions of interest. Experimental results on knee-oriented benchmark problems show that our algorithm outperforms state-of-the-art methods, with aggregation-dominance surpassing five existing knee identification techniques. We also validate the algorithm’s performance on the portfolio allocation problem.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"92 ","pages":"Article 101813"},"PeriodicalIF":8.2,"publicationDate":"2025-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143183499","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}

引用次数: 0

Q-learning assisted multi-objective evolutionary optimization for low-carbon scheduling of open-pit mine trucks

IF 8.2 1区计算机科学

Swarm and Evolutionary Computation Pub Date : 2025-02-01 DOI: 10.1016/j.swevo.2024.101778

Yao Huang , Yinan Guo , Guoyu Chen , Hong Wei , Xiaoxiao Zhao , Shengxiang Yang , Shirong Ge

{"title":"Q-learning assisted multi-objective evolutionary optimization for low-carbon scheduling of open-pit mine trucks","authors":"Yao Huang , Yinan Guo , Guoyu Chen , Hong Wei , Xiaoxiao Zhao , Shengxiang Yang , Shirong Ge","doi":"10.1016/j.swevo.2024.101778","DOIUrl":"10.1016/j.swevo.2024.101778","url":null,"abstract":"<div><div>Mine trucks, as the core equipment of discontinuous open-pit mining technology, account for high transportation costs and vast quantities of greenhouse gases. In order to improve transportation efficiency and decrease carbon emissions, rationally scheduling shovel-truck pairs is a necessary issue. Previous studies give less consideration on carbon emissions of trucks that varies with road and driving conditions. To overcome the shortage, a constraint bi-objective optimization model is built for low-carbon scheduling problem of open-pit mine trucks, in which minimizing both idle time and carbon emissions of trucks are taken as the objectives. More especially, the limits on working time, traffic volume and the number of trucks are modeled as the constraints. Carbon emissions is formulated by multistage nonlinear function that takes road condition, load and driving state of trucks into account. As the problem-solver, Q-learning assisted multi-objective evolutionary algorithm is put forward. Four evolution states are defined by analyzing the improvement on feasibility and convergence of the population, and four problem-specific evolution operators are designed to meet different demands of the evolution. Q-learning-based selection strategy is proposed to select the most appropriate operator, with the purpose of improving the evolution efficiency. Experimental results on the real-world instances expose that the proposed algorithm outperforms the other state-of-the-art algorithms significantly.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"92 ","pages":"Article 101778"},"PeriodicalIF":8.2,"publicationDate":"2025-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143183627","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}

引用次数: 0

DEEM — Differential Evolution with Elitism and Multi-populations

IF 8.2 1区计算机科学

Swarm and Evolutionary Computation Pub Date : 2025-02-01 DOI: 10.1016/j.swevo.2024.101818

Jan Machaček, Simon Siegel, Hauke Zachert

{"title":"DEEM — Differential Evolution with Elitism and Multi-populations","authors":"Jan Machaček, Simon Siegel, Hauke Zachert","doi":"10.1016/j.swevo.2024.101818","DOIUrl":"10.1016/j.swevo.2024.101818","url":null,"abstract":"<div><div>This paper introduces DEEM (Differential Evolution with Elitism and Multi-populations), a novel heuristic optimisation algorithm of the Differential Evolution family. DEEM integrates elitism and multi-population strategies to improve convergence speed and accuracy. Additionally, a diversity-based restart strategy is employed to significantly reduce the algorithm’s susceptibility to being trapped in local minima. The influence of algorithm parameter choices on optimisation success is demonstrated through a sensitivity study. The algorithm’s effectiveness is validated against benchmark functions from CEC 2015, 2017, 2020, and 2022, showing superior performance compared to state-of-the-art DE algorithms. Additionally, DEEM’s application is showcased through a complex optimisation problem in the field of geotechnical engineering: the calibration of advanced constitutive models for predicting the stress–strain behaviour of soils under monotonic and cyclic loading. This calibration process is notably time-consuming. DEEM not only achieves better objective values but also does so in fewer iterations, thus significantly reducing computational time.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"92 ","pages":"Article 101818"},"PeriodicalIF":8.2,"publicationDate":"2025-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143183507","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Adaptive search based Grey Wolf optimization algorithm for multi-objective optimization of ethylene cracking furnace

IF 8.2 1区计算机科学

Swarm and Evolutionary Computation Pub Date : 2025-02-01 DOI: 10.1016/j.swevo.2024.101810

Zhiqiang Geng , Weikang Kong , Xintian Wang , Ling Wang , Yongming Han

{"title":"Adaptive search based Grey Wolf optimization algorithm for multi-objective optimization of ethylene cracking furnace","authors":"Zhiqiang Geng , Weikang Kong , Xintian Wang , Ling Wang , Yongming Han","doi":"10.1016/j.swevo.2024.101810","DOIUrl":"10.1016/j.swevo.2024.101810","url":null,"abstract":"<div><div>The ethylene cracking furnace (ECF) is an important device for producing ethylene and propylene, so the optimization problem of the ECF is crucial. However, traditional optimization algorithms such as the grey wolf optimization (GWO) algorithm, are prone to getting stuck in local optima under the early stages and have low optimization accuracy under the later stage, which cannot effectively optimize the production of the ECF. Therefore, a novel multi-objective grey wolf optimization algorithm based on the adaptive search (ASMOGWO) is proposed. The non-linear convergence factor of the cosine transform in the ASMOGWO algorithm offsets its discovery and development capabilities. Then, the velocity formula of the GWO is updated based on the velocity update, effectively preventing individuals from entering local optima and improving the convergence performance. Meanwhile, the linearly decreasing inertia weight coefficients is proposed to control the convergence speed of the ASMOGWO. Compared with other optimization algorithms through public experiments, the ASMOGWO has good effects. Finally, the ASMOGWO algorithm is applied to optimize the ethylene yield and the propylene yield of the ECF. The result shows the proposed ASMOGWO has better feasibility than the original GWO algorithm and other optimization algorithms. Meanwhile, the optimized ethylene yield increased by 1.3570 %, while the propylene yield decreased by 0.0093 %.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"92 ","pages":"Article 101810"},"PeriodicalIF":8.2,"publicationDate":"2025-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143183597","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}

引用次数: 0

Learning memetic algorithm based on variable population and neighborhood for multi-complex target scheduling of large-scale imaging satellites

IF 8.2 1区计算机科学

Swarm and Evolutionary Computation Pub Date : 2025-02-01 DOI: 10.1016/j.swevo.2024.101789

Lei Li , Yonghao Du , Feng Yao , Shilong Xu , Yucheng She

{"title":"Learning memetic algorithm based on variable population and neighborhood for multi-complex target scheduling of large-scale imaging satellites","authors":"Lei Li , Yonghao Du , Feng Yao , Shilong Xu , Yucheng She","doi":"10.1016/j.swevo.2024.101789","DOIUrl":"10.1016/j.swevo.2024.101789","url":null,"abstract":"<div><div>The continual expansion of the scope and depth of satellite applications has resulted in the emergence of multiple complex targets, bringing significant challenges for imaging satellite task scheduling. To address the multicomplex target scheduling problem of large-scale heterogeneous imaging satellites (MCTS-LSIS), this article proposes a learning memetic algorithm based on variable population and neighborhood (LMA-VP/N). LMA-VP/N employs dual populations to achieve co-evolution and consists of a learning hybrid-rule heuristic, a collaboration and competition strategy, and a deep Q-network (DQN)-based variable neighborhood search (VNS) algorithm. Two encoding methods are designed to generate heterogeneous populations to explore different solution spaces. The learning hybrid-rule heuristic is proposed to produce high-quality individuals during initialization and diversity enhancement. The rule selection strategy is updated periodically based on elite knowledge. The collaboration and competition strategy is adopted to adaptively adjust the computing resources for the dual populations based on their performance. Moreover, eight VNS operators are developed based on two neighborhood structures and four neighborhood change rules. The DQN is adopted to select the proper action according to the state features of the populations. The experimental results indicate that LMA-VP/N outperforms both the state-of-the-art algorithms and conventional algorithms. The effectiveness of the main components of LMA-VP/N has also been verified.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"92 ","pages":"Article 101789"},"PeriodicalIF":8.2,"publicationDate":"2025-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143183599","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}

引用次数: 0

Promising boundaries explore and resource allocation evolutionary algorithm for constrained multiobjective optimization

IF 8.2 1区计算机科学

Swarm and Evolutionary Computation Pub Date : 2025-02-01 DOI: 10.1016/j.swevo.2024.101819

Yuelin Qu , Yuhang Hu , Wei Li , Ying Huang

{"title":"Promising boundaries explore and resource allocation evolutionary algorithm for constrained multiobjective optimization","authors":"Yuelin Qu , Yuhang Hu , Wei Li , Ying Huang","doi":"10.1016/j.swevo.2024.101819","DOIUrl":"10.1016/j.swevo.2024.101819","url":null,"abstract":"<div><div>Constrained multiobjective optimization problems (CMOPs) typically present numerous local optima, which can be deceptive. Current constrained multiobjective algorithms (CMOEAs) encounter challenges in maintaining diversity and escaping these local optima because of the single function of the population in the same space–time. Because they cannot keep exploring diversity and cannot balance their exploration focus. To this end, a dual-stage and dual-population algorithm named BPRRA is proposed in this article. Specifically, BPRRA utilizes new techniques to explore promising boundaries and allocate computing resources. In the first stage, one of the populations evolves to explore one promising boundary by ignoring constraints, and the other population explores another promising boundary by considering constraints. In the second stage, the two populations explore different regions from different promising boundaries using the diversity archiving strategy. Moreover, a novel resource allocation strategy is designed to dynamically allocate limited computational resources based on the ratio of potential offspring. The experiments involve five test suites and nine real-world problems to validate the performance of the proposed method. The results demonstrate that BPRRA has superior performance and can better solve CMOPs.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"92 ","pages":"Article 101819"},"PeriodicalIF":8.2,"publicationDate":"2025-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143183606","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}

引用次数: 0

Differential Evolution with multi-stage parameter adaptation and diversity enhancement mechanism for numerical optimization

IF 8.2 1区计算机科学

Swarm and Evolutionary Computation Pub Date : 2025-02-01 DOI: 10.1016/j.swevo.2024.101829

Qiutong Xu, Zhenyu Meng

{"title":"Differential Evolution with multi-stage parameter adaptation and diversity enhancement mechanism for numerical optimization","authors":"Qiutong Xu, Zhenyu Meng","doi":"10.1016/j.swevo.2024.101829","DOIUrl":"10.1016/j.swevo.2024.101829","url":null,"abstract":"<div><div>Differential Evolution (DE) is a powerful meta-heuristic algorithm for numerical optimization, however, it faces challenges such as improper parameter control, premature convergence, and population stagnation in complex problems. To address these issues, this paper proposes a Differential Evolution algorithm with multi-stage parameter adaptation and diversity enhancement mechanism (MD-DE). First, a multi-stage parameter adaptation scheme is designed, incorporating wavelet basis functions and Laplace distributions for parameter generation, and guiding parameter adjustment through a progressive Minkowski distance weighting strategy to balance exploration and exploitation. Second, a mutation strategy with dynamic dual archives is proposed, integrating potential information from promising but discarded solutions to enhance the diversity of donor vectors, thereby improving the perception of the fitness landscape. Finally, a hypervolume-based diversity metric is combined with a stagnation tracker to capture stagnant individuals, and a hierarchical intervention mechanism is employed to introduce perturbations, thereby enhancing the level of population diversity. To evaluate the performance of the proposed MD-DE, it was validated against five state-of-the-art DE variants on 87 benchmark functions from CEC2013, CEC2014, and CEC2017, as well as on real-world problems from CEC2011 and planetary gear design optimization problems. Experimental results demonstrate that our algorithm exhibits a high level of competitiveness.</div></div>","PeriodicalId":48682,"journal":{"name":"Swarm and Evolutionary Computation","volume":"92 ","pages":"Article 101829"},"PeriodicalIF":8.2,"publicationDate":"2025-02-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143183612","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}

引用次数: 0