Evolutionary Algorithm Research Articles

Managing uncertainty in production sequencing: a digital twin framework for mixed-model assembly lines

ABSTRACT In response to dynamic and uncertain contexts, companies adopt mixed-model lines over single lines to ensure flexibility and meet diverse, low-volume production needs. Consequently, sequencing plays a central role in multi-objective daily production optimization. To identify optimal solutions, the multi-objective sequencing problem needs metaheuristics, especially evolutionary approaches, while digital twins manage uncertainty in dynamic production contexts. This work proposes a digital twin framework combining evolutionary algorithms and simulation to optimize sequencing for mixed-model lines under stochastic conditions. Validation through simulation confirms its feasibility and effectiveness. A case study in a leather goods company validates the framework’s applicability in the fashion industry, showcasing managerial benefits for decision-makers. Additionally, its iterative, data-driven implementation ensures easy updates, crucial for High Variety/Low Volume environments. Academically, the study innovates by combining evolutionary algorithms and simulation, demonstrating their synergy in real-world optimization and providing a replicable approach for companies operating in highly variable production scenarios.

Evolution algorithm with adaptive genetic operator and dynamic scoring mechanism for large-scale sparse many-objective optimization

Large-scale sparse multi-objective optimization problems are prevalent in numerous real-world scenarios, such as neural network training, sparse regression, pattern mining and critical node detection, where Pareto optimal solutions exhibit sparse characteristics. Ordinary large-scale multi-objective optimization algorithms implement undifferentiated update operations on all decision variables, which reduces search efficiency, so the Pareto solutions obtained by the algorithms fail to meet the sparsity requirements. SparseEA is capable of generating sparse solutions and calculating scores for each decision variable, which serves as a basis for crossover and mutation in subsequent evolutionary process. However, the scores remain unchanged in iterative process, which restricts the sparse optimization ability of the algorithm. To solve the problem, this paper proposes an evolution algorithm with the adaptive genetic operator and dynamic scoring mechanism for large-scale sparse many-objective optimization (SparseEA-AGDS). Within the evolutionary algorithm for large-scale Sparse (SparseEA) framework, the proposed adaptive genetic operator and dynamic scoring mechanism adaptively adjust the probability of cross-mutation operations based on the fluctuating non-dominated layer levels of individuals, concurrently updating the scores of decision variables to encourage superior individuals to gain additional genetic opportunities. Moreover, to augment the algorithm’s capability to handle many-objective problems, a reference point-based environmental selection strategy is incorporated. Comparative experimental results demonstrate that the SparseEA-AGDS algorithm outperforms five other algorithms in terms of convergence and diversity on the SMOP benchmark problem set with many-objective and also yields superior sparse Pareto optimal solutions.

Elemental numerical descriptions to enhance classification and regression model performance for high-entropy alloys

The machine learning-assisted design of new alloy compositions often relies on the physical and chemical properties of elements to describe the materials. In the present study, we propose a strategy based on an evolutionary algorithm to generate new elemental numerical descriptions for high-entropy alloys (HEAs). These newly defined descriptions significantly enhance classification accuracy, increasing it from 77% to ~97% for recognizing FCC, BCC, and dual phases, compared to traditional empirical features. Our experimental validation demonstrates that our classification model, utilizing these new elemental numerical descriptions, successfully predicted the phases of 8 out of 9 randomly selected alloys, outperforming the same model based on traditional empirical features, which correctly predicted 4 out of 9. By incorporating these descriptions derived from a simple logistic regression model, the performance of various classifiers improved by at least 15%. Moreover, these new numerical descriptions for phase classification can be directly applied to regression model predictions of HEAs, reducing the error by 22% and improving the R2 value from 0.79 to 0.88 in hardness prediction. Testing on six different materials datasets, including ceramics and functional alloys, demonstrated that the obtained numerical descriptions achieved higher prediction precision across various properties, indicating the broad applicability of our strategy.

Halving the cost of quantum algorithms with randomization

Quantum signal processing (QSP) provides a systematic framework for implementing a polynomial transformation of a linear operator, and unifies nearly all known quantum algorithms. In parallel, recent works have developed randomized compiling, a technique that promotes a unitary gate to a quantum channel and enables a quadratic suppression of error (i.e., ϵ → O(ϵ2)) at little to no overhead. Here we integrate randomized compiling into QSP through Stochastic Quantum Signal Processing. Our algorithm implements a probabilistic mixture of polynomials, strategically chosen so that the average evolution converges to that of a target function, with an error quadratically smaller than that of an equivalent individual polynomial. Because nearly all QSP-based algorithms exhibit query complexities scaling as O(log(1/ϵ))—stemming from a result in functional analysis—this error suppression reduces their query complexity by a factor that asymptotically approaches 1/2. By the unifying capabilities of QSP, this reduction extends broadly to quantum algorithms, which we demonstrate on algorithms for real and imaginary time evolution, phase estimation, ground state preparation, and matrix inversion.

Reinforcement learning for mutation operator selection in automated program repair

Automated program repair techniques aim to aid software developers with the challenging task of fixing bugs. In heuristic-based program repair, a search space of mutated program variants is explored to find potential patches for bugs. Most commonly, every selection of a mutation operator during search is performed uniformly at random, which can generate many buggy, even uncompilable programs. Our goal is to reduce the generation of variants that do not compile or break intended functionality which waste considerable resources. In this paper, we investigate the feasibility of a reinforcement learning-based approach for the selection of mutation operators in heuristic-based program repair. Our proposed approach is programming language, granularity-level, and search strategy agnostic and allows for easy augmentation into existing heuristic-based repair tools. We conducted an extensive empirical evaluation of four operator selection techniques, two reward types, two credit assignment strategies, two integration methods, and three sets of mutation operators using 30,080 independent repair attempts. We evaluated our approach on 353 real-world bugs from the Defects4J benchmark. The reinforcement learning-based mutation operator selection results in a higher number of test-passing variants, but does not exhibit a noticeable improvement in the number of bugs patched in comparison with the baseline, uniform random selection. While reinforcement learning has been previously shown to be successful in improving the search of evolutionary algorithms, often used in heuristic-based program repair, it has yet to demonstrate such improvements when applied to this area of research.

Multi-objective optimization of composite kayak paddle shaft through evolutionary deep learning

ABSTRACT This paper presents a methodology for multiobjective optimization of a kayak paddle shaft made of Carbon Fiber Reinforced Polymer (CFRP) composite by a filament winding process for the case of individualization of the structure for a single sportperson. Design parameters of a composite layup (orientation of fibers and thickness) and objective functions of the paddle shaft (deflection, strength and weight) are described. These objective functions are based on the individual kayaker’s performance and preferences. Data driven evolutionary algorithms, aided by mechanics based theory and experimentation, were used for deep learning and optimization. Due to the mutually contradicting objective functions, the result of such optimization is a Pareto-optimal solution set for an individual sportsperson from which a preferred solution can be picked up by a decision maker. This work shows that such methodology could be successfully used for a personalization of any sport equipment.

A Two-Stage Multi-Objective Optimization Algorithm for Solving Large-Scale Optimization Problems

For large-scale multi-objective optimization, it is particularly challenging for evolutionary algorithms to converge to the Pareto Front. Most existing multi-objective evolutionary algorithms (MOEAs) handle convergence and diversity in a mutually dependent manner during the evolution process. In this case, the performance degradation of one solution may lead to the deterioration of the performance of the other solution. This paper proposes a two-stage multi-objective optimization algorithm based on decision variable clustering (LSMOEA-VT) to solve large-scale optimization problems. In LSMOEA-VT, decision variables are divided into two categories and use dimensionality reduction methods to optimize the variables that affect evolutionary convergence. Following this, we performed an interdependence analysis to break down the convergence variables into multiple subcomponents that are more tractable. Furthermore, a non-dominated dynamic weight aggregation method is used to enhance the diversity of the population. To evaluate the performance of our proposed algorithm, we performed extensive comparative experiments against four optimization algorithms across a diverse set of benchmark problems, including eight multi-objective optimization problems and nine large-scale optimization problems. The experimental results show that our proposed algorithm performs well on some test functions and has certain competitiveness.

Creating Effective Self-Adaptive Differential Evolution Algorithms to Solve the Discount-Guaranteed Ridesharing Problem Based on a Saying

Sustainable transport is an important trend in smart cities to achieve sustainability development goals. It refers to the use of transport modes with low emissions, energy consumption and negative impacts on the environment. Ridesharing is one important sustainable transport mode to attain the goal of net zero greenhouse gas emissions. The discount-guaranteed ridesharing problem (DGRP) aims to incentivize drivers and riders and promote ridesharing through the guarantee of a discount. However, the computational complexity of the DGRP poses a challenge in the development of effective solvers. In this study, we will study the effectiveness of creating new self-adaptive differential evolution (DE) algorithms based on an old saying to solve the DGRP. Many old sayings still have far-reaching implications today. Some of them influence the organization of management teams for companies and decisions to improve performance and efficiency. Whether an old saying that works effectively for human beings to solve problems can also work for developers to create effective optimization problem solvers in the realm of artificial intelligence is an interesting research question. In our previous study, one self-adaptive algorithm was proposed to solve the DGRP. This study demonstrates how to create a series of self-adaptive algorithms based on the old saying “Two heads are better than one” and validates the effectiveness of this approach based on experiments and comparison with the algorithms proposed previously. The new finding of this study is that the old saying not only works effectively for human beings to solve problems but also works effectively in the creation of new scalable and robust self-adaptive algorithms to solve the DGRP. In other words, the old saying provides a simple and systematic approach to the development of effective optimization problem solvers in artificial intelligence.

A buffer allocation evolutionary algorithm for resource-constrained projects with activity clusters

A buffer allocation evolutionary algorithm for resource-constrained projects with activity clusters

Multi-Layer LEO Constellation Optimization Based on D-NSDE Algorithm

Low-Earth-orbit (LEO) satellites have unique advantages in communication, navigation, and remote sensing due to their low orbit, strong landing signal strength, and low launch cost. However, the optimization of the design of LEO constellations to obtain the optimal configuration to meet different missions faces great challenges. Traditional multi-objective optimization algorithms often struggle with designing constellations involving composite functions due to various constraints, which can result in premature termination and local optimality issues. This paper introduces a dynamic parameter-based non-dominated sorting differential evolution (D-NSDE) algorithm to obtian better solutions, which is capable of dynamically adjusting the boundary of feasible solutions and modifying operators according to the iteration process to mitigate these constraints. Additionally, we model a composite LEO constellation with multiple layers, constructing 2-/3-/4-layer configurations, and we include constraints from the third-generation BeiDou Navigation Satellite System (BDS-3) navigation constellations. Subsequently, we employ the D-NSDE algorithm to solve the corresponding multi-objective optimization problems and derive the optimal solution set. The results demonstrate that D-NSDE can generate complete and multi-level solution sets under diverse constraint conditions, with 75% of D-NSDE algorithm optimization solutions being able to achieve seamless positioning for 95% of global coverage. Furthermore, the PDOP median values are 5.12/4.23/2.97 without BDS-3 navigation constraints and 1.38/1.44/1.51 with BDS-3 navigation constraints. Additionally, simulation experiments conducted on standard function test sets confirm that the solution sets produced by the D-NSDE algorithm exhibit favorable distribution and convergence performance better than the Non-dominated Sorting Genetic Algorithm (NSGA)-III.

Predicting the diversity of photosynthetic light-harvesting using thermodynamics and machine learning.

Oxygenic photosynthesis is responsible for nearly all biomass production on Earth, and may have been a prerequisite for establishing a complex biosphere rich in multicellular life. Life on Earth has evolved to perform photosynthesis in a wide range of light environments, but with a common basic architecture of a light-harvesting antenna system coupled to a photochemical reaction centre. Using a generalized thermodynamic model of light-harvesting, coupled with an evolutionary algorithm, we predict the type of light-harvesting structures that might evolve in light of different intensities and spectral profiles. We reproduce qualitatively the pigment composition, linear absorption profile and structural topology of the antenna systems of multiple types of oxygenic photoautotrophs, suggesting that the same physical principles underlie the development of distinct antenna structures in various light environments. Finally we apply our model to representative light environments that would exist on Earth-like exoplanets, predicting that both oxygenic and anoxygenic photosynthesis could evolve around low mass stars, though the latter would seem to work better around the coolest M-dwarfs. We see this as an interesting first step toward a general evolutionary model of basic biological processes and proof that it is meaningful to hypothesize on the nature of biology beyond Earth.

Integrating neural network and multimedia technologies to enhance college students’ career development

Combining neural network technologies and computational techniques, this research establishes a career development promotion system based on a multi-modal neural network. It reveals that computer simulation technology and multimedia have positive intervention effects on college students’ career decision-making behaviors, similar to how biomolecular interactions regulate biological processes. This technology ensures scientific rigor, objectivity, and authenticity. A knowledge fusion algorithm, built on attributes and rules within the Hadoop platform and MapReduce parallel computing framework, facilitates effective data integration. Additionally, inspired by the regulatory mechanisms in biomolecular systems, a neural network-based algorithm, utilizing gradient descent, is applied to cultural learning, augmented by feedback analysis to assess students’ psychological changes, posture, and response dynamics during the learning process. To further optimize the career development framework, an Evolutionary Algorithm (EA) is used to enhance the performance of neural networks. Numerical simulations demonstrate the robustness of the proposed algorithm, achieving high accuracy (0.981), recall rate (1.0), and F-measure (0.997) in similarity computations. These results are particularly notable when biomechanic metrics, such as gesture and posture tracking, are integrated with linguistic data, such as spelling and vocabulary. The findings underscore that incorporating neural network insights into multimedia teaching methodologies can significantly enhance psychological motivation, behavioral adaptability, and engagement in college students, fostering improved educational outcomes and advancing interdisciplinary innovation in neural networks. It effectively enhances the internal driving force of “technology empowering psychological development” in the career planning system and provides a cognitive computing and biomechanic perspective for the construction of the smart education ecosystem.

Gender recognition in text-independent speaker identification using MFCC, spectrogram, Bi-LSTM, and rat swarm evolutionary algorithm optimization

Gender recognition in text-independent speaker identification using MFCC, spectrogram, Bi-LSTM, and rat swarm evolutionary algorithm optimization

SedInterFoam 1.0: a three-phase numerical model for sediment transport applications with free surfaces

Abstract. In this paper, an Eulerian two-phase flow model, sedFoam, is extended to include an air phase together with its water and sediment phases. The numerical model called sedInterFoam is implemented using the open-source library OpenFOAM. sedInterFoam includes the previous features of sedFoam for sediment transport modeling and also solves the air–water interface using the volume-of-fluid method coupled with the waves2Foam toolbox for free-surface wave generation and absorption. Using sedInterFoam, four test cases are successfully reproduced to validate the free-surface evolution algorithm's implementation, mass conservation of sediment and fluid phases, and predictive capabilities and to demonstrate its potential in modeling a broader range of coastal applications with sediment transport dominated by surface waves.

Fitness and Collaborative Information-Driven Differential Evolution Algorithm for Bound-Constrained and Real-World Application Problems

Fitness and Collaborative Information-Driven Differential Evolution Algorithm for Bound-Constrained and Real-World Application Problems

3D spatial evolutionary particle swarm algorithm based emergency communication spatial deployment optimization

ABSTRACT Building emergency communication in disaster areas is a key problem that emergency response needs to solve. Ad hoc Network (ANET) can quickly establish communication networks when public communication infrastructure is disrupted. At present, ground emergency communication deployment based on ANET usually relies on the operator’s experience, which struggles to ensure high-quality deployment in complex urban environments. This paper proposes an ANET nodes spatial optimization deployment algorithm based on 3D Spatial Evolutionary Particle Swarm Optimization (3DSEPSO). A wireless communication transmission rate model that depends on surface buildings and trees is constructed using ground truth communication data. The algorithm incorporates a fitness evaluation model, particle chromosome structure, and an evolutionary mechanism to intelligently deploy ANET nodes. By considering the spatial distribution of buildings and trees, the algorithm can achieve optimal data transmission quality by using a given number of ANET nodes. Experimental results demonstrate that the proposed algorithm significantly outperforms empirical approaches and traditional methods in terms of data transmission rates and quality. Thus, the algorithm provides better support for emergency rescue teams by facilitating more effective and reliable emergency communication.

CrySPAI: A New Crystal Structure Prediction Software Based on Artificial Intelligence

Crystal structure predictions based on the combination of first-principles calculations and machine learning have achieved significant success in materials science. However, most of these approaches are limited to predicting specific systems, which hinders their application to unknown or unexplored domains. In this paper, we present a crystal structure prediction software based on artificial intelligence, named as CrySPAI, to predict energetically stable crystal structures of inorganic materials given their chemical compositions. The software consists of three key modules, an evolutionary optimization algorithm (EOA) that searches for all possible crystal structure configurations, density functional theory (DFT) that provides the accurate energy values for these structures, and a deep neural network (DNN) that learns the relationship between crystal structures and their corresponding energies. To optimize the process across these modules, a distributed framework is implemented to parallelize tasks, and an automated workflow has been integrated into CrySPAI for seamless execution. This paper reports the development and implementation of the AI-based CrySPAI Crystal Prediction Software tool and its unique features.

A multi‐objective feature optimization strategy for developing high‐entropy alloys with optimal strength and ductility

AbstractSelecting appropriate material features is essential for effective data‐driven materials design. Here, we propose a multi‐objective feature optimization strategy that identifies feature subsets to improve both prediction accuracy and active learning efficiency for iterative experimentation. Our approach integrates an evolutionary genetic algorithm to explore an expanded feature space, encompassing both traditional feature pools and a continuous numerical representation of elements rather than relying solely on discrete values. We demonstrate this strategy by identifying high‐entropy alloys (HEAs) with optimal strength and ductility. Results show that the optimized feature subsets reduce prediction errors by 20% for strength and 11% for ductility. Additionally, within fewer than three feedback iterations, HEAs with outstanding combinations of yield strength and ductility are identified, highlighting the high efficiency of this approach. This multi‐objective feature optimization strategy is adaptable to other material systems, offering a pathway to improve machine learning performance and accelerate materials discovery.

Energy-efficient cluster head selection in Internet of Things networks using an optimized evaporation rate water-cycle algorithm

This paper presents a new scheme for energy-efficient clustering in Internet of Things (IoT) networks by employing an optimized evolutionary rate water cycle algorithm (OERWCA), aiming to address crucial issues, such as energy conservation measured through average energy consumption per node, network longevity quantified by total operational rounds until node depletion, and throughput as an indicator of data transmission efficiency. In OERWCA, a local escaping operator (LEO) is introduced to avoid algorithm trapping in local optima by enhancing its exploration capability. Besides, advanced control-randomization operators balance exploration and exploitation dynamically for efficient search behavior in the solution space. The algorithm optimizes cluster head selection by minimizing energy consumption and redundant transmission. Simulations comparing OERWCA with previous optimization methods, including NCCLA, FHHO, and EACH-COA, demonstrate the superior performance of the proposed algorithm. Key metrics evaluated include network lifetime, throughput, average transmission delay, packet delivery ratio (PDR), and energy efficiency. OERWCA achieves significant improvements, including up to a 26% increase in network lifetime, a 32% boost in throughput, a 20% reduction in transmission delay, and a 27% enhancement in PDR compared to the best-performing benchmarks. These results highlight OERWCA’s effectiveness in optimizing critical performance parameters for IoT networks. The enhanced convergence properties of the proposed algorithm also address some common limitations found in existing methods. This work, therefore, provides a robust solution toward extending the operational lifetime of IoT networks, which is one of the fundamental steps forward in large-scale efficient resource management.

A Discriminative Prediction Strategy Based on Multi-View Knowledge Transfer for Dynamic Multi-Objective Optimization

Dynamic multi-objective optimization problems (DMOPs) are widely encountered in engineering optimization processes, characterized by conflicting objectives that change over time. Evolutionary transfer optimization (ETO) has recently emerged as a promising optimization paradigm for addressing DMOPs. ETO-based dynamic multi-objective evolutionary algorithms (DMOEAs) usually adopt discriminative prediction strategies to select high-quality initial solutions in new environments by transferring knowledge from the source domain, thereby accelerating the evolutionary process. However, DMOPs often change in both decision and objective spaces. Existing methods rely on a single view from the source domain for knowledge transfer, which can overlook crucial knowledge from other views, potentially affecting the selection of high-quality initial solutions and limiting optimization performance. To address this, we propose a discriminative prediction strategy based on multi-view knowledge transfer for DMOEA, called MKT-DMOEA. Specifically, we construct discriminative predictors from the view of the decision space and objective space in the source domain. Each discriminative predictor effectively reduces the differences between environments under the current view through a transfer strategy based on geometric feature transformation. Meanwhile, these predictors make full use of the transferred knowledge to achieve accurate discriminative predictions under the current view. Finally, we integrate the discriminative prediction results from multiple views to select high-quality initial solutions. Experimental results demonstrate that our proposed algorithm outperforms four other state-of-the-art DMOEAs in terms of both diversity and convergence on the well-known CEC 2018 dynamic multi-objective optimization benchmark suite DF.