IEEE Transactions on Evolutionary Computation最新文献_第4页

A Coevolutionary Response Framework for Dynamic Constrained Multiobjective Optimization Problems 动态约束多目标优化问题的协同进化响应框架

IF 15.9 1区计算机科学

IEEE Transactions on Evolutionary Computation Pub Date : 2026-10-01 Epub Date: 2025-08-04 DOI: 10.1109/TEVC.2025.3595410

Qian Bao;Maocai Wang;Shengxiang Yang;Guangming Dai;Xiaoyu Chen

{"title":"A Coevolutionary Response Framework for Dynamic Constrained Multiobjective Optimization Problems","authors":"Qian Bao;Maocai Wang;Shengxiang Yang;Guangming Dai;Xiaoyu Chen","doi":"10.1109/TEVC.2025.3595410","DOIUrl":"10.1109/TEVC.2025.3595410","url":null,"abstract":"Dynamic constrained multiobjective optimization problems (DCMOPs) present significant challenges due to the evolving nature of both objectives and constraints. These problems require optimization algorithms that can efficiently adapt to dynamic environments while maintaining a balance between convergence and diversity. To address these challenges, we propose a novel cooperative response dynamic constrained multiobjective optimization (CRDCMO) framework. The framework introduces two key strategies: 1) population reinitialization guided by historical environmental information, tailored to different types of environmental changes and 2) dynamic adjustment of auxiliary population tasks, optimizing resource allocation with a focus on tracking the constrained Pareto fronts (CPF). These strategies enhance the algorithm’s adaptability to environmental changes and improve CPF tracking efficiency. The CRDCMO framework is extensively evaluated on several benchmark test suites, as well as a real-world energy optimization problem. Experimental results demonstrate that CRDCMO outperforms seven state-of-the-art algorithms, underscoring its effectiveness and robustness in dynamic environments. This framework not only provides a comprehensive solution for DCMOPs but also contributes to the advancement of dynamic optimization algorithms.","PeriodicalId":13206,"journal":{"name":"IEEE Transactions on Evolutionary Computation","volume":"30 5","pages":"1803-1817"},"PeriodicalIF":15.9,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144778342","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

Parametric Pareto Set Learning: Amortizing Multiobjective Optimization With Parameters 参数Pareto集学习：带参数的平摊多目标优化

IF 15.9 1区计算机科学

IEEE Transactions on Evolutionary Computation Pub Date : 2026-10-01 Epub Date: 2025-11-10 DOI: 10.1109/TEVC.2025.3630787

Ji Cheng;Xi Lin;Bo Xue;Qingfu Zhang

{"title":"Parametric Pareto Set Learning: Amortizing Multiobjective Optimization With Parameters","authors":"Ji Cheng;Xi Lin;Bo Xue;Qingfu Zhang","doi":"10.1109/TEVC.2025.3630787","DOIUrl":"10.1109/TEVC.2025.3630787","url":null,"abstract":"Parametric multiobjective optimization (PMO) addresses the challenge of solving an infinite number of multiobjective optimization problems (MOPs), where optimal solutions must adapt to varying parameters. Traditional methods, such as parameterized multiobjective evolutionary algorithms (MOEAs), generate only a finite set of solutions and are unable to capture the continuous structure of the Pareto set (PS) across the entire parameter space, thereby leaving critical tradeoffs unexplored. To overcome these challenges, we propose parametric Pareto Set Learning (PPSL), a novel framework that leverages amortized optimization to learn a unified mapping from parameters to the PS with respect to the parameter. The framework employs a hypernetwork conditioned on input parameters, enabling real-time inference of optimal solutions for any parameter setting without the need for repeated optimization. By integrating low-rank adaptation (LoRA), parametric PSL (PPSL) achieves computational efficiency and scalability for PMO problems. To verify its effectiveness, we apply PPSL to dynamic MOPs (DMOPs) and MOPs with shared component design, where the parameters encode time-dependent or modular constraints. The experimental results demonstrate that PPSL significantly outperforms existing methods in terms of solution quality and adaptability. Importantly, we emphasize that PPSL is not limited to the two problems tested in this work; it is broadly applicable to any real-world problem that can be modeled as a PMO problem, offering a versatile and powerful solution for a wide range of applications. Code is available at: <uri>https://github.com/jicheng9617/PPSL</uri>.","PeriodicalId":13206,"journal":{"name":"IEEE Transactions on Evolutionary Computation","volume":"30 5","pages":"2054-2067"},"PeriodicalIF":15.9,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145484979","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

Estimation of Hitting Time by Hitting Probability for Elitist Evolutionary Algorithms 基于命中概率的精英进化算法命中时间估计

IF 15.9 1区计算机科学

IEEE Transactions on Evolutionary Computation Pub Date : 2026-10-01 Epub Date: 2025-11-12 DOI: 10.1109/TEVC.2025.3632072

Jun He;Siang Yew Chong;Xin Yao

{"title":"Estimation of Hitting Time by Hitting Probability for Elitist Evolutionary Algorithms","authors":"Jun He;Siang Yew Chong;Xin Yao","doi":"10.1109/TEVC.2025.3632072","DOIUrl":"10.1109/TEVC.2025.3632072","url":null,"abstract":"Drift analysis is one of the main tools for analyzing the time complexity of evolutionary algorithms (EAs). However, it requires manual construction of drift functions to bound hitting time for each specific algorithm and problem. To address this limitation, linear drift functions were introduced for elitist EAs. But calculating good linear bound coefficients remains a problem. This article proposes a new method called drift analysis of hitting probability to compute these coefficients. Each coefficient is interpreted as a bound on the hitting probability of a fitness level, transforming the task of estimating hitting time into estimating hitting probability. A new drift analysis method is then developed to estimate hitting probability, where paths are introduced to handle multimodal fitness landscapes. Explicit expressions are constructed to compute hitting probability, significantly simplifying the estimation process. An advantage of the proposed method is its ability to estimate both the lower and upper bounds of hitting time and to compare the performance of two algorithms in terms of hitting time. To demonstrate this application, two algorithms for the knapsack problem, each incorporating feasibility rules and greedy repair, respectively, are compared. The analysis indicates that neither constraint handling technique consistently outperforms the other.","PeriodicalId":13206,"journal":{"name":"IEEE Transactions on Evolutionary Computation","volume":"30 5","pages":"2011-2023"},"PeriodicalIF":15.9,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145498391","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

GPLight+: A Genetic Programming Method for Learning Symmetric Traffic Signal Control Policy 一种学习对称交通信号控制策略的遗传规划方法

IF 15.9 1区计算机科学

IEEE Transactions on Evolutionary Computation Pub Date : 2026-10-01 Epub Date: 2025-06-11 DOI: 10.1109/TEVC.2025.3578575

Xiao-Cheng Liao;Yi Mei;Mengjie Zhang

{"title":"GPLight+: A Genetic Programming Method for Learning Symmetric Traffic Signal Control Policy","authors":"Xiao-Cheng Liao;Yi Mei;Mengjie Zhang","doi":"10.1109/TEVC.2025.3578575","DOIUrl":"10.1109/TEVC.2025.3578575","url":null,"abstract":"Recently, learning-based approaches have achieved significant success in automatically devising effective traffic signal control (TSC) strategies. In particular, as a powerful evolutionary machine learning approach, genetic programming (GP) is utilized to evolve human-understandable phase urgency functions to measure the urgency of activating a green light for a specific phase. However, current GP-based methods are unable to treat the common traffic features of different traffic signal phases consistently. To address this issue, we propose to use a symmetric phase urgency function to calculate the phase urgency for a specific phase based on the current road conditions. This is represented as an aggregation of two shared subtrees, each representing the urgency of a turn movement in the phase. We then propose a GP method to evolve the symmetric phase urgency function. We evaluate our proposed method on the well-known cityflow traffic simulator, based on multiple public real-world datasets. The experimental results show that the proposed symmetric urgency function representation can significantly improve the performance of the learned TSC policies over the traditional GP representation on a wide range of scenarios. Further analysis shows that the proposed method can evolve effective, human-understandable, and easily deployable TSC policies.","PeriodicalId":13206,"journal":{"name":"IEEE Transactions on Evolutionary Computation","volume":"30 5","pages":"1818-1832"},"PeriodicalIF":15.9,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144268612","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

Expensive Constrained Optimization via Adaptive Switching Between Surrogate-Assisted Constrained and Unconstrained Search 基于代理辅助约束搜索与无约束搜索自适应切换的昂贵约束优化

IF 15.9 1区计算机科学

IEEE Transactions on Evolutionary Computation Pub Date : 2026-10-01 Epub Date: 2025-11-06 DOI: 10.1109/TEVC.2025.3629639

Yuanchao Liu;Jianchang Liu;Lixin Tang

{"title":"Expensive Constrained Optimization via Adaptive Switching Between Surrogate-Assisted Constrained and Unconstrained Search","authors":"Yuanchao Liu;Jianchang Liu;Lixin Tang","doi":"10.1109/TEVC.2025.3629639","DOIUrl":"10.1109/TEVC.2025.3629639","url":null,"abstract":"expensive constrained optimization problems (ECOPs) are prevalent in many engineering domains, where evaluating objective and constraints requires costly simulations or physical experiments. As powerful optimization methods in artificial intelligence, surrogate-assisted evolutionary algorithms (SAEAs) have demonstrated effectiveness in solving ECOPs. However, most existing SAEAs focus solely on surrogate-assisted constrained search, overlooking the potential benefits offered by surrogate-assisted unconstrained search. Therefore, a surrogate-assisted differential evolution with adaptive switching between constrained and unconstrained search is proposed for expensive constrained optimization. The proposed method aims to integrate both surrogate-assisted constrained and unconstrained search modes within a unified framework, enabling adaptive switching between the two search modes during the optimization process. Moreover, a selection ratio update mechanism is proposed to dynamically adjust the selection ratio between the two search modes in accordance with their contributions to algorithm performance. Experimental results on test suites and a real-world engineering application demonstrate that the proposed method outperforms five state-of-the-art SAEAs in solving ECOPs.","PeriodicalId":13206,"journal":{"name":"IEEE Transactions on Evolutionary Computation","volume":"30 5","pages":"1981-1995"},"PeriodicalIF":15.9,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145455279","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 Generative Adversarial Network Based Prediction Strategy for Evolutionary Dynamic Multiobjective Optimization 基于生成对抗网络的进化动态多目标优化预测策略

IF 15.9 1区计算机科学

IEEE Transactions on Evolutionary Computation Pub Date : 2026-10-01 Epub Date: 2025-09-17 DOI: 10.1109/TEVC.2025.3611014

Feng Wang;Jinsong Xie;Fanshu Liao;Yixuan Li;Yinan Guo;Shengxiang Yang;Aimin Zhou

{"title":"A Generative Adversarial Network Based Prediction Strategy for Evolutionary Dynamic Multiobjective Optimization","authors":"Feng Wang;Jinsong Xie;Fanshu Liao;Yixuan Li;Yinan Guo;Shengxiang Yang;Aimin Zhou","doi":"10.1109/TEVC.2025.3611014","DOIUrl":"10.1109/TEVC.2025.3611014","url":null,"abstract":"Recently, dynamic multiobjective evolutionary optimization has attracted much research attention, and the prediction based strategy has been proved to be an effective way for dynamic multiobjective optimization evolutionary algorithms (dynamic multiobjective optimization evolutionary algorithm) to track the changing Pareto set (PS). Recently, some strategies have been proposed to use the information from both the historical and new environments to predict the population in the new environment. However, their performance in complex new environments is limited, as they partially utilize information from the new environment. generative adversarial networks (GANs) have been proven to be an efficient generative model, which can capture the relationships between high-quality solutions in the new environment and estimate the data distribution more accurately. In this article, we propose a GAN-based prediction strategy (GANPS) to generate more high-quality solutions that can adapt to the new environment. In GANPS, a PS estimation method is employed to improve the quality of samples (solutions) for GAN training at first, which helps the GAN better extract the information from historical and new environment. Then GANPS utilizes the trained generator to reinitialize the population, and a model reuse strategy is further designed to reduce training costs and the algorithm’s response time. Experimental results on a set of benchmark test suites show that GANPS could outperform other state-of-the-art reaction strategies in most cases and has better change tracking capability in dynamic environments.","PeriodicalId":13206,"journal":{"name":"IEEE Transactions on Evolutionary Computation","volume":"30 5","pages":"1878-1892"},"PeriodicalIF":15.9,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145077689","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 Crowding Distance That Provably Solves the Difficulties of the NSGA-II in Many-Objective Optimization 可证明地解决NSGA-II多目标优化困难的拥挤距离

IF 15.9 1区计算机科学

IEEE Transactions on Evolutionary Computation Pub Date : 2026-10-01 Epub Date: 2025-11-28 DOI: 10.1109/TEVC.2025.3638306

Weijie Zheng;Yan Gao;Benjamin Doerr

{"title":"A Crowding Distance That Provably Solves the Difficulties of the NSGA-II in Many-Objective Optimization","authors":"Weijie Zheng;Yan Gao;Benjamin Doerr","doi":"10.1109/TEVC.2025.3638306","DOIUrl":"10.1109/TEVC.2025.3638306","url":null,"abstract":"Recent mathematical runtime analyses have shown that the crowding distance used by the nondominated sorting genetic algorithm II (NSGA-II) multiobjective optimizer does not well reflect how isolated, and hence worth keeping, solutions are. This has provably led to drastic difficulties when the number of objectives is three or more. Based on the insights of these works, we design a variant of the crowding distance, called truthful crowding distance, and analyze the resulting truthful NSGA-II with mathematical means: 1) we show that it solves the many-objective versions of theOMM, COCZ, LOTZ, and <inline-formula> <tex-math>${text {OJZJ}}_{k}$ </tex-math></inline-formula> problems in the same (polynomial) asymptotic runtimes as the NSGA-III and the SMS-EMOA, contrasting the exponential lower bounds for the classic NSGA-II; 2) for the bi-objective versions of these problems, our NSGA-II maintains the good performance of the classic NSGA-II, but is successful already with population sizes smaller by a constant factor; and 3) for the bi-objective OMM problem, we observe a (minimally) better performance in approximating the Pareto front. These results indicate that our truthful crowding distance is an interesting alternative to the classic crowding distance.","PeriodicalId":13206,"journal":{"name":"IEEE Transactions on Evolutionary Computation","volume":"30 5","pages":"2082-2092"},"PeriodicalIF":15.9,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145613439","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 Multicriteria Game-Driven Constrained Multiobjective Evolutionary Algorithm 多准则博弈驱动约束多目标进化算法

IF 15.9 1区计算机科学

IEEE Transactions on Evolutionary Computation Pub Date : 2026-10-01 Epub Date: 2025-12-08 DOI: 10.1109/TEVC.2025.3641368

Heming Jia;Yongchao Li;Hongguang Li

{"title":"A Multicriteria Game-Driven Constrained Multiobjective Evolutionary Algorithm","authors":"Heming Jia;Yongchao Li;Hongguang Li","doi":"10.1109/TEVC.2025.3641368","DOIUrl":"10.1109/TEVC.2025.3641368","url":null,"abstract":"Solving constrained multiobjective optimization problems (CMOPs) involves simultaneously achieving convergence to the Pareto front, preserving solution diversity, and satisfying complex constraints. To address these challenges, this article proposes a multicriteria game-driven constrained multiobjective evolutionary algorithm (MGCMOEA), which leverages the complementary roles of a main population and an auxiliary population to jointly explore both feasible and infeasible regions. The algorithm introduces a refined feasibility classification mechanism that differentiates weakly feasible solutions from infeasible ones, thereby accelerating convergence while preserving potentially valuable solutions. To enhance cooperation between populations, MGCMOEA incorporates a multicriteria game-theoretic selection model that integrates dynamic game theory with the technique for order preference by similarity to ideal solution (TOPSIS). This model improves the guidance provided by the auxiliary population during evolution. Furthermore, to overcome inefficiencies associated with static population sizes, a dynamic resource allocation strategy is adopted, enabling adaptive distribution of computational resources throughout the evolutionary process. Experimental results on 47 benchmark functions and 12 real-world CMOPs demonstrate that MGCMOEA consistently outperforms 8 state-of-the-art algorithms in terms of convergence, diversity, and feasibility.","PeriodicalId":13206,"journal":{"name":"IEEE Transactions on Evolutionary Computation","volume":"30 5","pages":"2123-2137"},"PeriodicalIF":15.9,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145704000","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

LAS-GP: Harnessing Large Semantic Libraries for Evolutionary Feature Construction in Symbolic Regression 利用大型语义库构建符号回归中的进化特征

IF 15.9 1区计算机科学

IEEE Transactions on Evolutionary Computation Pub Date : 2026-10-01 Epub Date: 2025-11-14 DOI: 10.1109/TEVC.2025.3632839

Hengzhe Zhang;Qi Chen;Bing Xue;Wolfgang Banzhaf;Mengjie Zhang

{"title":"LAS-GP: Harnessing Large Semantic Libraries for Evolutionary Feature Construction in Symbolic Regression","authors":"Hengzhe Zhang;Qi Chen;Bing Xue;Wolfgang Banzhaf;Mengjie Zhang","doi":"10.1109/TEVC.2025.3632839","DOIUrl":"10.1109/TEVC.2025.3632839","url":null,"abstract":"Evolutionary feature construction has seen notable advancements in recent years. A major limitation affecting the effectiveness of evolutionary feature construction is its failure to fully leverage the large number of building blocks evaluated during the evolutionary process. In this article, to fully utilize historical building blocks, we propose a semantic local search algorithm that operates in the semantic space using a dynamically updated semantic library. This library collects up to 10 000 trees and subtrees during the evolutionary process, with a design that adapts linear scaling and considers tree complexity to enhance both accuracy and generalization performance. To manage a large semantic library, we employ a ball tree and an instance selection mechanism, optimizing computational and memory resources. Furthermore, a positive–negative search mechanism and a semantic library-guided mutation operator are introduced to further improve the effectiveness of the semantic library in feature construction. Experimental validation on 120 datasets demonstrates that our semantic library-based genetic programming approach significantly outperforms several symbolic regression and machine learning algorithms, highlighting its superior performance.","PeriodicalId":13206,"journal":{"name":"IEEE Transactions on Evolutionary Computation","volume":"30 5","pages":"2039-2053"},"PeriodicalIF":15.9,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145515613","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

IEEE Transactions on Evolutionary Computation Publication Information IEEE进化计算汇刊信息

IF 15.9 1区计算机科学

IEEE Transactions on Evolutionary Computation Pub Date : 2026-10-01 Epub Date: 2026-10-02 DOI: 10.1109/TEVC.2026.3730924

引用次数: 0