Han Li, Dario Landa-Silva · Evolutionary Computation · 2011
A multi-objective optimization problem can be solved by decomposing it into one or more single objective subproblems in some multi-objective metaheuristic algorithms. Each subproblem corresponds to one weighted aggregation function. For example, MOEA/D is an evolutionary multi-objective optimization (EMO) algorithm that attempts to optimize multiple subproblems simultaneously by evolving a population of solutions. However, the performance of MOEA/D highly depends on the initial setting and diversity of the weight vectors. In this paper, we present an improved version of MOEA/D, called EMOSA, which incorporates an advanced local search technique (simulated annealing) and adapts the search directions (weight vectors) corresponding to various subproblems. In EMOSA, the weight vector of each subproblem is adaptively modified at the lowest temperature in order to diversify the search toward the unexplored parts of the Pareto-optimal front. Our computational results show that EMOSA outperforms six other well established multi-objective metaheuristic algorithms on both the (constrained) multi-objective knapsack problem and the (unconstrained) multi-objective traveling salesman problem. Moreover, the effects of the main algorithmic components and parameter sensitivities on the search performance of EMOSA are experimentally investigated.
Cited by 176Open access
Xiangyu Wang · Journal of Engineering Mechanics and Machinery · 2025
To address the multi-performance conflict in lightweight robotic arm design, this study establishes a systematic design framework integrating multi-objective optimization. By combining topology optimization with parametric modeling to develop lightweight structural solutions, and applying advanced optimization algorithms to resolve multi-objective coordination issues involving mass, stiffness, and dynamic characteristics, the framework generates a Pareto solution set that characterizes performance trade-offs. A comprehensive performance evaluation system is further developed to assess the synergy between static-dynamic characteristics and energy efficiency precision, validating the optimization outcomes. The research demonstrates that this framework effectively guides designers in achieving optimal balance between structural lightweighting and overall performance under multiple constraints, providing a methodological basis for high-performance robotic arm development that combines theoretical rigor with engineering applicability.
Cited by 0Open access
H. Li · 2014
A multi-objective optimization problem can be solved by decomposing it into one or more single objective subproblems in some multi-objective metaheuristic algorithms. Each subproblem corresponds to one weighted aggregation function. For example, MOEA/D is an evolutionary multi-objective optimization (EMO) algorithm that at-tempts to optimize multiple subproblems simultaneously by evolving a population of solutions. However, the performance of MOEA/D highly depends on the initial setting and diversity of the weight vectors. In this paper, we present an improved version of MOEA/D, called EMOSA, which incorporates an advanced local search technique (simulated annealing) and adapts the search directions (weight vectors) cor-responding to various subproblems. In EMOSA, the weight vector of each subprob-lem is adaptively modified at the lowest temperature in order to diversify the search towards the unexplored parts of the Pareto-optimal front. Our computational results show that EMOSA outperforms six other well-established multi-objective metaheuris-tic algorithms on both the (constrained) multi-objective knapsack problem and the (un-constrained) multi-objective traveling salesman problem. Moreover, the effects of the main algorithmic components and parameter sensitivities on the search performance of EMOSA are experimentally investigated.
Cited by 0
Wentao Zhang · 2025
Urban last-mile logistics distribution faces significant challenges due to dynamic factors such as real-time traffic conditions, order fluctuations, and delivery constraints. This paper presents a comprehensive study on dynamic path planning models and algorithms tailored for urban last-mile logistics. A multi-objective mathematical model is developed to minimize delivery time, cost, and environmental impact while considering vehicle capacity, time windows, and real-time data integration. Three advanced algorithms-Dynamic Dijkstra with Real-Time Weight Adjustment (DDRWA), Reinforcement Learning-Enhanced Ant Colony Optimization (RL-ACO), and Multi-Objective Particle Swarm Optimization (MOPSO)-are proposed and evaluated. Experimental results using the Solomon benchmark dataset and real-world logistics data demonstrate that the proposed algorithms outperform traditional methods in terms of path efficiency and adaptability to dynamic changes.
Cited by 0
Ha Minh Hieu, Hoa Phan, Tung Duy Doan et al. · Proceedings of the AAAI Conference on Artificial Intelligence · 2026
Multi-objective combinatorial optimization problems (MOCOP) frequently arise in practical applications that require the simultaneous optimization of conflicting objectives. Although traditional evolutionary algorithms can be effective, they typically depend on domain knowledge and repeated parameter tuning, limiting flexibility when applied to unseen MOCOP instances. Recently, integration of Large Language Models (LLMs) into evolutionary computation has opened new avenues for automatic heuristic generation, using their advanced language understanding and code synthesis capabilities. Nevertheless, most existing approaches predominantly focus on single-objective tasks, often neglecting key considerations such as runtime efficiency and heuristic diversity in multi-objective settings. To bridge this gap, we introduce Multi-heuristics for MOCOP via Pareto-Grid-guided Evolution of LLMs (MPaGE), a novel enhancement of the Simple Evolutionary Multiobjective Optimization (SEMO) framework that leverages LLMs and Pareto Front Grid (PFG) technique. By partitioning the objective space into grids and retaining top-performing candidates to guide heuristic generation, MPaGE utilizes LLMs to prioritize heuristics with semantically distinct logical structures during variation, thus promoting diversity and mitigating redundancy within the population. Through extensive evaluations, MPaGE demonstrates superior performance over existing LLM-based frameworks, and achieves competitive results to traditional Multi-objective evolutionary algorithms (MOEAs), with significantly faster runtime.
Cited by 0Open access
Ayyoub W. Al Hourani, M. A. Abido, Mahmoud Kassas · IEEE Access · 2025
Transformers are essential components of electrical power systems, representing a significant capital investment and requiring optimized designs to balance efficiency, cost, and performance. As the demand for energy-efficient and cost-effective solutions increases, transformer manufacturers face increasing pressure to enhance design efficiency while minimizing production costs. Traditional deterministic optimization techniques often struggle with the high-dimensional, nonlinear, and multi-objective nature of transformer design problems, necessitating the adoption of advanced evolutionary optimization strategies. In recent years, evolutionary algorithms (EAs) have emerged as powerful tools for addressing these challenges, offering adaptability, scalability, and robustness in optimizing transformer designs. This paper presents a comprehensive review of transformer design optimization using EAs, investigating both single and multi-objective approaches applied to distribution, power, and other transformer types. A wide range of publications are reviewed, categorized, and analyzed to provide a structured overview of recent advancements in Genetic Algorithms (GA), Differential Evolution (DE), Particle Swarm Optimization (PSO), and hybrid AI-assisted approaches. This study highlights key research gaps, including oversimplification of transformer models, limited experimental validation, and insufficient focus on multi-objective optimization. Future directions emphasize the integration of finite element analysis (FEA) into optimization loops, AI-assisted optimization frameworks, and sustainability-oriented transformer designs. These insights provide valuable guidance for both researchers and practitioners, contributing to the advancement of cost-effective, high-efficiency, and industry-applicable transformer design methodologies using evolutionary algorithms.
Cited by 1Open access
Trung Hieu Tran, Quoc-Anh Vu, Viet-Hung Truong et al. · Civil Engineering Journal · 2024
This study enriches structural optimization research using direct analysis for steel truss structures, which is often hampered by high computational demands. The main objective of this work is to evaluate multi-objective optimization algorithms in truss sizing optimization with discrete variables, focusing on minimizing total mass and controlling inter-story drift under multiple load combinations. Five leading multi-objective metaheuristic algorithms were assessed: SPEA2, GDE3, NSGA2, MOEA/D, and the novel MOEA/D-EpDE, which uniquely combines MOEA/D with Dynamical Resource Allocation and pbest Differential Evolution. Four performance indicators, such as Generational Distance (GD), GD Plus (GD+), Inverted GD+ (IGD+), and Hypervolume (HV), were utilized. Findings from four truss optimization problems revealed that all considered algorithms located feasible optimal solutions, but MOEA/D-EpDE excelled, consistently securing the lowest GD, GD+, IGD+, and anchor point values, along with the highest HV values in most scenarios. This indicates its superior capability in addressing the problem efficiently. NSGA2 and MOEA/D also performed well, outperforming GDE3 and SPEA2. This study is pioneering in its application of these algorithms to steel truss optimization via direct analysis, highlighting the potential for advanced computational techniques in structural engineering. Doi: 10.28991/CEJ-2024-010-12-07 Full Text: PDF
Cited by 1Open access
Joshua Odion, Chukwu F. Nelson, Aniekan B. Inyang et al. · Zenodo (CERN European Organization for Nuclear Research) · 2023
Interest on FACTS controllers in the scientific and engineering community have increased due to their high reliability, low cost, and absence of infrastructural overhauling. In the context of global sustainability, it is required that FACTS controllers must fulfill physical, technological, and environmental requirements, through design and optimization studies. This study provides a comprehensive evaluation of the state-of-the-art technologies relevant to multi-objective optimization and performance analysis of FACTS controllers. Types, general overview, FACTS deployed in the Nigerian power systems and multi-objective location optimization and their impact were discussed. Several efficient optimization strategies from recent research works are highlighted and compared, showing that there are research gaps on the integration of advanced control algorithms and use of robust optimization methods.
Cited by 0Open access
Joshua Odion, Chukwu F. Nelson, Aniekan B. Inyang et al. · INTERNATIONAL JOURNAL OF MULTIDISCIPLINARY RESEARCH AND ANALYSIS · 2023
Interest on FACTS controllers in the scientific and engineering community have increased due to their high reliability, low cost, and absence of infrastructural overhauling. In the context of global sustainability, it is required that FACTS controllers must fulfill physical, technological, and environmental requirements, through design and optimization studies. This study provides a comprehensive evaluation of the state-of-the-art technologies relevant to multi-objective optimization and performance analysis of FACTS controllers. Types, general overview, FACTS deployed in the Nigerian power systems and multi-objective location optimization and their impact were discussed. Several efficient optimization strategies from recent research works are highlighted and compared, showing that there are research gaps on the integration of advanced control algorithms and use of robust optimization methods.
Cited by 2Open access
Yuexi Peng, Zhiyong Huang, Zhijun Li et al. · IEEE Internet of Things Journal · 2025
With the swift rise in the number of digital images, their transmission and storage are facing severe challenges. Although traditional encryption algorithms can effectively protect image privacy, the noise-like features they generate may attract more attention from attackers. To address these problems, a new visually meaningful image encryption algorithm based on compressive sensing and a multi-objective optimization method is proposed. The proposed algorithm includes an improvement method for measurement matrix based on singular value decomposition and an advanced embedding method, and a multi-objective optimization algorithm is used for parameter selection. Simulation results show that the proposed algorithm performs well in reconstruction quality, visual security, and efficiency. Compared with existing algorithms, the proposed method demonstrates significant advantages in efficiency, reconstruction quality, embedding rate, and robustness.
Cited by 2
Ravipudi Venkata Rao, Joao Paulo Davim · Metals · 2025
This paper presents three simple and efficient advanced optimization algorithms, namely the best–worst–random (BWR), best–mean–random (BMR), and best–mean–worst–random (BMWR) algorithms designed to address unconstrained and constrained single- and multi-objective optimization tasks of the metal casting processes. The effectiveness of the algorithms is demonstrated through real case studies, including (i) optimization of a lost foam casting process for producing a fifth wheel coupling shell from EN-GJS-400-18 ductile iron, (ii) optimization of process parameters of die casting of A360 Al-alloy, (iii) optimization of wear rate in AA7178 alloy reinforced with nano-SiC particles fabricated via the stir-casting process, (iv) two-objectives optimization of a low-pressure casting process using a sand mold for producing A356 engine block, and (v) four-objectives optimization of a squeeze casting process for LM20 material. Results demonstrate that the proposed algorithms consistently achieve faster convergence, superior solution quality, and reduced function evaluations compared to simulation software (ProCAST, CAE, and FEA) and established metaheuristics (ABC, Rao-1, PSO, NSGA-II, and GA). For single-objective problems, BWR, BMR, and BMWR yield nearly identical solutions, whereas in multi-objective tasks, their behaviors diverge, offering well-distributed Pareto fronts and improved convergence. These findings establish BWR, BMR, and BMWR as efficient and robust optimizers, positioning them as promising decision support tools for industrial metal casting.
Cited by 4Open access
Arash Heidari, Sebastian Rojas Gonzalez, Ivo Couckuyt · Proceedings of the Genetic and Evolutionary Computation Conference Companion · 2025
Multi-objective optimization algorithms aim to identify a set of non-dominated solutions that capture the complex and often conflicting trade-offs between objectives. In black-box settings, where the input-output relationship is usually analytically intractable, these algorithms face two major challenges: limited evaluation budgets, especially when objective evaluations are costly, and computational difficulties in proposing new candidates. While all non-dominated solutions hold equal mathematical value, the Decision-Maker is typically interested in only the most preferred trade-offs. This adds an additional challenge for preference-driven optimization: the cost of human interaction. To address these challenges, this paper proposes leveraging surrogate-assisted evolutionary multi-objective optimization methods, combined with advanced discretization techniques and strategic filtering of preferred solutions to mitigate the overall computational burden. After evaluating against similar methods in the literature, the experimental results suggest the method to be competitive in enhancing both the efficiency and precision of the decision-making process.
Cited by 0Open access
Christian Pérez, Carlos March, Miguel Á. Salido · Proceedings of the Genetic and Evolutionary Computation Conference Companion · 2025
This paper introduces IGJSP, a Benchmark Generator for Job Shop Scheduling Problems (JSP), which generates customized instances to evaluate optimization algorithms. JSP requires assigning tasks to jobs and machines while minimizing makespan, energy consumption, and tardiness, key aspects for industrial process optimization and energy-efficient scheduling. IGJSP fills a gap by providing flexible, reproducible benchmarks that mirror real-world complexities, unlike existing libraries such as JSPLIB that do not support modern multi-objective scenarios essential for advanced optimization techniques.
Cited by 0
Zahra Saeidi Mobarakeh, Hossein Amoozadkhalili · SHILAP Revista de lepidopterología · 2024
This research introduces a nonlinear multi-objective optimization model that is designed to simultaneously optimize profit and customer satisfaction in production systems. The investigated problem includes optimization in complex and uncertain conditions of production, which is faced with resource and time limitations. The proposed model provides optimal solutions for managers by using non-linear objective functions and detailed analysis of operating conditions. This fuzzy logic is combined with machine learning algorithms such as neural networks and reinforcement learning to create an intelligent and flexible model that effectively adapts to sudden changes in dynamic environments. This model uses the combination of non-dominant fourth sorting genetic algorithms (NSGA-IV) and variable selection network (VSN) in a hybrid framework and provides an advanced and multi-faceted approach to solving complex multi-objective optimization problems. Pareto-optimal results obtained from this model indicate its efficient and optimal performance. The proposed model can be used as a practical and strategic source for managers and decision makers in optimizing production and improving customer satisfaction in uncertain and dynamic conditions.
Cited by 0Open access
Richard Yi Zhang, John G. Kassakian, Richard Yi Zhang · 2012
This thesis describes an efficient numerical simulation technique of magnetoquasistatic electromagnetic fields for planar induction heating applications. The technique is based on a volume-element discretization, integral formulation of Maxwell’s equa-tions, and uses the multilayer Green’s function to avoid volumetric meshing of the heated material. The technique demonstrates two orders of magnitude of computa-tional advantage compared to existing FEM techniques. Single-objective and multi-objective optimization of a domestic induction heating coil are performed using the new technique, using more advanced algorithms than those previously used due to the increase in speed. Both optimization algorithms produced novel, three-dimensional induction coil designs.
Cited by 0
Minh Ha, Hung M. Phan, Therese Doan et al. · ArXiv.org · 2025
Multi-objective combinatorial optimization problems (MOCOP) frequently arise in practical applications that require the simultaneous optimization of conflicting objectives. Although traditional evolutionary algorithms can be effective, they typically depend on domain knowledge and repeated parameter tuning, limiting flexibility when applied to unseen MOCOP instances. Recently, integration of Large Language Models (LLMs) into evolutionary computation has opened new avenues for automatic heuristic generation, using their advanced language understanding and code synthesis capabilities. Nevertheless, most existing approaches predominantly focus on single-objective tasks, often neglecting key considerations such as runtime efficiency and heuristic diversity in multi-objective settings. To bridge this gap, we introduce Multi-heuristics for MOCOP via Pareto-Grid-guided Evolution of LLMs (MPaGE), a novel enhancement of the Simple Evolutionary Multiobjective Optimization (SEMO) framework that leverages LLMs and Pareto Front Grid (PFG) technique. By partitioning the objective space into grids and retaining top-performing candidates to guide heuristic generation, MPaGE utilizes LLMs to prioritize heuristics with semantically distinct logical structures during variation, thus promoting diversity and mitigating redundancy within the population. Through extensive evaluations, MPaGE demonstrates superior performance over existing LLM-based frameworks, and achieves competitive results to traditional Multi-objective evolutionary algorithms (MOEAs), with significantly faster runtime. Our code is available at: https://github.com/langkhachhoha/MPaGE.
Cited by 0Open access
Mehmet Fatih Karakaş, Fatma Lati̇foğlu · International Journal of Advances in Signal and Image Sciences · 2025
Digital signal processing in biomedical applications requires advanced filtering techniques that can simultaneously optimize multiple performance criteria. This study introduces a novel multi-objective metaheuristic approach for Finite Impulse Response (FIR) filter design, addressing the complex challenges of signal processing optimization. The research employs advanced metaheuristic algorithms, including Particle Swarm Optimization (PSO), Artificial Bee Colony (ABC), Genetic Algorithm (GA), Black Widow Optimization (BWO), Chaos Game Optimization (CGO), Harmony Search (HS), Squirrel Search Algorithm (SSA), and Atomic Orbital Search (AOS) to optimize multiple filter design parameters jointly. In contrast to conventional approaches, the proposed multi-objective optimization strategy demonstrates superior performance in signal filtering. The proposed method considers the important design trade-offs at the same time, such as minimizing signal power after stopband frequency, reducing stopband ripple, maximizing stopband first lobe and stopband edge frequency attenuation, and reducing computational complexity. The method is employed on ElectroCardioGram (ECG) signals from the MIT-BIH open-access database. Performance comparison indicated that multi-objective metaheuristic filters achieved much better performance than conventional FIR filters. The optimized filters exhibited improved stopband attenuation, narrower transition bands, lower power leakage, and more efficient signal processing at a cut-off frequency of 100 Hz. Signal power measurements demonstrated significant improvements. While conventional FIR filters ranged from -37.2001 dB to -41.46778 dB, multi-objective metaheuristic filters reached -42.05318 dB to -44.69498 dB in terms of stopband power.
Cited by 1Open access
Haoming Zhang, Qianlong Dang, Xinyu Feng et al. · Swarm and Evolutionary Computation · 2025
Solving constrained multi-objective optimization problems requires simultaneously satisfying multiple objectives and constraints, presenting a significant challenge for solving tasks. Constructing auxiliary optimization problems to assist the main optimization problem in accelerating convergence is a common approach for constrained multi-objective evolutionary algorithms (CMOEAs). However, this approach may waste computational resources on auxiliary optimization problems that provide little benefit to the main optimization problem at certain stages of the evolution process. Based on the above issue, this paper proposes a constrained multi-objective optimization algorithm to address this issue via competitive mechanism based reward auxiliary optimization problems (RACMO). Specifically, an unconstrained auxiliary optimization problem and a dynamic constrained auxiliary optimization problem are constructed. They are rewarded by the number of solutions provided to the main optimization problem, and the cumulative reward is mapped to the probability to adaptively selecting more valuable auxiliary optimization problems. Moreover, an adaptive stop-update strategy is designed. By controlling the competition between two auxiliary populations and adaptive stop-updating, excellent convergence is guaranteed while significantly saving computational resources. Experimental results demonstrate the competitiveness of RACMO compared to ten advanced CMOEAs on three test suites and eight practical application problems.
Cited by 3Open access
Fangpei Yao, Gai‐Ge Wang · Applied Sciences · 2023
Dynamic multi-objective optimization problems (DMOPs) have become a research hotspot in engineering optimization, because their objective functions, constraints, or parameters may change over time, while quickly and accurately tracking the changing Pareto optimal set (POS) during the optimization process. Therefore, solving dynamic multi-objective optimization problems presents great challenges. In recent years, transfer learning has been proved to be one of the effective means to solve dynamic multi-objective optimization problems. However, this paper proposes a new transfer learning method based on clustering difference to solve DMOPs (TCD-DMOEA). Different from the existing methods, it uses the clustering difference strategy to optimize the population quality and reduce the data difference between the target domain and the source domain. On this basis, transfer learning technology is used to accelerate the construction of initialization population. The advantage of the TCD-DMOEA method is that it reduces the possibility of negative transfer and improves the performance of the algorithm by improving the similarity between the source domain and the target domain. Experimental results show that compared with several advanced dynamic multi-objective optimization algorithms based on different benchmark problems, the proposed TCD-DMOEA method can significantly improve the quality of the solution and the convergence speed.
Cited by 9Open access
Tao Ma, Jun Zhong, Fang Xu et al. · Journal of Industrial and Management Optimization · 2025
In this paper, a Multi-objective Distributed Heterogeneous No-Idle Flowshop Scheduling Problem with Sequence-Dependent Setup Time (MDHNIFSP-SDST) is studied. Firstly, a multi-objective optimization model was established with the objective of minimizing makespan (completion time) and Total Tardiness (TT). According to the characteristics of the problem, a Discrete Sine Optimization Algorithm (DSOA) is proposed. This algorithm mainly includes four core stages: In the multi-neighborhood search stage, a search strategy based on key factories was proposed, and the Q-learning algorithm was used to enable individuals to select appropriate operators during evolution; In the stage of destruction and reconstruction, an iterative search strategy was designed to guide the evolutionary direction of individuals, while the sinusoidal optimization algorithm was used to balance global development and local search. In the selection stage, the method of non-dominated sorting and reference points is used to screen for high-quality solutions, and the external file set is used to store all non-dominated solutions. In the collaborative stage, the sharing and competition mechanisms among populations are designed to balance the optimization of the two objectives. Finally, 270 instances were solved by DSOA algorithm and advanced algorithms at the same time and measured by multi-objective indicators. The results show that DSOA algorithm ranks first in all test instances compared with other algorithms, which verifies the effectiveness of DSOA algorithm.
Cited by 0Open access