Differential Evolution Algorithm Research Articles

A differential evolution framework based on the fluid model for feature selection

Feature selection in machine learning is a crucial step to effectively address the issue of feature redundancy in classification problems. Numerous feature selection algorithms have been developed to minimize the number of features, reduce computational cost, and improve classification accuracy. Differential evolution algorithms have the advantage of being simple in structure, robust, fast in convergence, and frequently used to solve feature selection problems. However, it is worth noting that differential evolution algorithms are susceptible to local optimum and stagnation issues, particularly when applied to high-dimensional data. To address this issue, in this study, we propose a differential evolution framework based on the fluid model, named DEF-FM, for feature selection. DEF-FM has the capability to speed up the convergence of differential evolution algorithms and alleviate the effects of local optima. The proposed framework is validated and compared against eight popular differential evolution algorithms using 12 publicly available benchmark datasets and experimental results unequivocally demonstrate the superiority of the proposed framework.

Dynamic optimization of die-less spinning process through autonomous modeling of time-varying forming law

Die-less spinning is a typical incremental forming technology with flexible processing parameters, including the roller path and loading parameters, whose forming law changes dynamically and easy-to-produce defects of excessive thinning and flange wrinkling. This brings a great challenge to the process design of die-less spinning. In this study, we proposed a novel dynamic optimization approach and system considering the time-varying forming law for the processing parameters in die-less spinning. The idea of proposed approach is that simulating the spinning process step by step through autonomous finite element modeling, real-time extracting and modeling the simulated spinning status at each forming step to capture the time-varying forming law, then gradually optimizing the processing parameters through efficient multi-objective optimization algorithm. By conducting the proposed dynamic optimization approach to the first spinning pass of die-less spinning, series of polynomial response surface models (PRSM) of wall thickness and flange wrinkling at various forming steps were developed to describe the time-varying forming law. On this basis, the roller path and matching loading parameters in each step were gradually optimized through a differential evolution algorithm with the multi-objective of minimizing wall thickness reduction, reducing flange wrinkling and maximizing forming efficiency. The optimized process produced better spun part than the traditional spinning process, presenting the wall thickness reduction ≤5 % and the flange wrinkling ≤2.5 mm. Furthermore, the time-varying influence of processing parameters on the wall thickness and flange wrinkling were elucidated based on the PRSM of spinning status at various forming steps.

Multi-User Detection Based on Improved Cheetah Optimization Algorithm

Targeting the issues of slow speed and inadequate precision of optimal solution calculation for multi-user detection in complex noise environments, this paper proposes a multi-user detection algorithm based on a Hybrid Cheetah Optimizer (HCO). The algorithm first optimizes the control parameters and individual update mechanism of the Cheetah Optimizer (CO) algorithm using a nonlinear strategy to improve the uniformity and discretization of the individual search range, and then dynamically introduces a differential evolutionary algorithm into the improved selection mechanism of the CO algorithm, which is utilized to fine-tune the solution space and maintain the local diversity during the fast search process. Simulation results demonstrate that this detection algorithm not only realizes fast convergence with a very low bit error rate (BER) at eight iterations but also has obvious advantages in terms of noise immunity, resistance to far and near effects, communication capacity, etc., which greatly improves the speed and accuracy of optimal position solving for multi-user detection and can achieve the purpose of accurate solving in complex environments.

A hybrid optimization algorithm to identify unknown parameters of photovoltaic models under varying operating conditions

Enhancing the performance of optimization algorithms to accurately extract parameters for photovoltaic modules is a significant challenge in the field. While several metaheuristic algorithms have been explored, achieving a balance between exploration and exploitation capabilities remains elusive. In this context, hybrid metaheuristics have emerged as a promising approach to address this issue and improve optimization efficiency. In our study, we propose a novel hybrid algorithm that integrates the backtracking search optimization algorithm (BSA) with the differential evolution (DE) algorithm. This hybridization aims to enhance the precision of parameter extraction for various photovoltaic models, including single-diode, dual-diode, and commercially available modules such as the ST40 Thin Film, SM55 Monocrystalline, and S75 Polycrystalline. The BSA is initially employed to leverage historical experiences and guide the evolutionary process, thereby enhancing the algorithm’s exploration capabilities. Subsequently, the DE algorithm is introduced to accelerate convergence during iterations and improve exploitation of the search space. By combining these two algorithms, we aim to achieve superior accuracy and efficiency in parameter estimation. To validate the effectiveness of our approach, extensive experiments are conducted, and the results demonstrate its superiority over other sophisticated algorithms. The proposed algorithm achieves remarkable performance, as evidenced by the lowest mean square error values obtained: 9.86021877891501E-04 for the single-diode model, 9.82484851785148E-04 for the dual-diode model, 2.42507486809511E-03 for Photowatt-PWP201, 1.72981370994069E-03 for STM6-40/36, and 1.66006031250855E-02 for STP6-120/36. This highlights the algorithm’s exceptional accuracy and efficiency in extracting parameters for photovoltaic modules.

Innovative integrated workflow for data-driven production forecasting and well completion optimization: A Montney Formation case study

With the use of advanced data analytics and machine learning (ML) techniques, the oil and gas industry has transformed significantly in recent years. These innovations have opened new opportunities in reservoir engineering, allowing engineers to leverage data for better decision-making and well completion optimization. A key challenge in hydrocarbon exploration and production is forecasting gas production rates from unconventional sources. This is a crucial factor that affects how resources are allocated and decisions are made, requiring the application of novel techniques to enhance hydrocarbon production. Here, a novel and comprehensive automated workflow is introduced covering data collection, data preparation, feature selection, ML model development, and operational parameter optimization. The workflow uses various ML and optimization methods, including a modified Differential Evolution (MDE) algorithm, to improve gas production forecasting and optimize well completion parameters. We applied the proposed workflow to 3144 horizontal gas wells in the Montney Formation, British Columbia (BC), Canada, to assess its performance. The results showed the advantages of incorporating the initial 24 months of cumulative gas production data alongside well and completion data. Also, following the fine-tuning of hyperparameters for the ML algorithms, we identified the most effective model for forecasting cumulative gas production within the Montney Formation as a two-layer Artificial Neural Network (ANN) model. Moreover, using the MDE algorithm, the optimization results revealed the possibility of increasing gas production by more than 10% for about 47% of the wells. This research makes a significant contribution to the field of unconventional gas exploration by introducing an automated workflow that simplifies data processing, model development, and decision-making processes. It also offers valuable insights for optimizing unconventional gas production in the Montney Formation and similar reservoirs, as the energy industry moves towards more sustainable and efficient practices.

Assessing Residential Building Energy Efficiency Using Evolutionary Dendritic Neural Regression

Assessing building energy consumption is of paramount significance in sustainability and energy efficiency (EE) studies. The development of an accurate EE prediction model is pivotal for optimizing energy resources and facilitating effective building planning. Traditional physical modeling approaches are encumbered by high complexity and protracted modeling cycles. In this paper, we introduce a novel evolutionary dendritic neural regression (EDNR) model tailored to forecasting residential building EE. Acknowledging the vast landscape and complexity of the EDNR weight space, coupled with the inherent susceptibility of traditional optimization algorithms to local optima, we propose a complex network-guided strategy-based differential evolution algorithm for training the EDNR model. This strategy adeptly strikes a balance between exploration and exploitation during the search process, significantly enhancing the predictive and generalization capacities of EDNR. To our knowledge, this study represents the inaugural application of dendritic neural regression in real-world prediction scenarios. Extensive experimental findings demonstrate the efficacy of EDNR in accurately predicting building EE with commendable performance. Furthermore, the results of two nonparametric statistical tests affirm the validity and stability of EDNR. Consequently, our proposed methodology exhibits high potential and competitiveness in machine learning applications within the energy domain.

Acceleration model considering multi‐stress coupling effect and reliability modeling method based on nonlinear Wiener process

AbstractEstablishing an accurate accelerated degradation model is paramount for ensuring precise reliability evaluation results. Unfortunately, current accelerated degradation tests often lack test groups for investigating multi‐stress coupled phenomena. Consequently, existing multi‐stress accelerated models fail to adequately consider the impact of stress coupling when data with stress coupling information is absent. This limitation leads to the development of inaccurate models, ultimately affecting the precision of reliability assessment. To address this challenge, this paper introduces a new modeling method for multi‐stress accelerated degradation models that takes into account stress coupling effects. The proposed modeling method aims to improve the accuracy of reliability assessment under multi‐stress conditions. In the proposed model, the main effect function of stress is determined based on existing single‐stress accelerated models. The coupling effect is first examined through the Multivariate Analysis of Variance (MANOVA), and then the functional form of the coupling effect function is determined from the given candidate functions through correlation analysis. Next, the coupling effect is incorporated into a Wiener process to establish a multi‐stress accelerated degradation model, and the two‐step estimation method combining Least Squares Method (LSM) and Differential Evolution Algorithm (DEA) is proposed. The accuracy and effectiveness of the coupling effect test method, model establishment, and parameter estimation method were validated using two Monte Carlo simulation experimental data sets. Finally, the superiority of the proposed model is demonstrated through examples, providing feasible ideas and technical support for the research on multi‐stress accelerated degradation modeling considering stress coupling.

A cost-effective seed selection model for multi-constraint influence maximization in social networks

The Influence Maximization (IM) problem aims to maximize the diffusion of information or adoption of products among users within a social network by identifying and activating a set of initial users. It is not unrealistic to have a higher activation cost for more influential users in real-life applications. However, existing works on IM focus solely on finding the most influential users as the seed set, without considering either the activation costs of individual nodes and the total budget or the size of the seed set. This oversight may lead to infeasible solutions, particularly from financial and managerial perspectives, respectively. To address these issues, we propose a more realistic and generalized formulation called Multi-Constraint Influence Maximization (MCIM) to achieve a cost-effective solution under both budgetary and cardinality constraints. The MCIM model allows for variable-length solutions, necessitating the exclusion of seed nodes from the influence spread estimation. Consequently, unlike existing IM formulations, the spread function under the MCIM model is no longer monotonic but submodular. As it has also been proven to be an NP-hard problem, we propose a Simple Additive Weighting (SAW)-assisted Differential Evolution (DE) algorithm for solving large-size real-world IM problems. Experimental results on four datasets demonstrate the effectiveness of the proposed formulation and algorithm in finding optimal and cost-effective solutions.

Configuration and Operation Optimization for Industrial Steam Power System Sustainable Retrofit Considering Renewable Energy Integration

AbstractRenewable energy integration and operational optimization are crucial in energy sustainability and decarbonization, especially for industrial steam power systems (SPS). This study establishes an SPS superstructure that integrates wind, solar, and biomass energy. A mixed‐integer nonlinear programming (MINLP) model is developed to determine an optimal retrofit strategy that minimizes the life cycle cost, which includes carbon emission cost and energy cost. To solve this high‐dimensional complex optimization problem, a multistage strategy fusion differential evolution algorithm with dynamic partitioning of the SVM feasible domain (SVM‐MS‐DE) is developed. The results from the optimal strategy demonstrate a 9.02% reduction in the system's total cost and a 9.64% decrease in carbon emission through the incorporation of wind and solar energy. Additionally, the sensitivity analysis on biomass ratio, carbon emission price, energy demand, and carbon emission limits reveal that the region can contribute significantly to renewable energy initiatives. Several recommended policies are provided to encourage enterprises to move towards sustainable development.

Efficient boundary conditions identification in thermal simulation of the spindle system with reduced order model and differential evolution algorithm

The boundary conditions in thermal simulation of the spindle system are very complicated and the empirically calculated ones will usually result in a significant discrepancy between experiment and simulation. Thus, the optimal boundary conditions were identified by treating it as an inverse optimization problem in this paper. Moreover, to advance the optimization efficiency, a reduced order model is established to replace the time-consuming thermal simulation of the spindle system. By superposition of truncated eigenmodes and accurately predicted modal coefficients, the reduced order model can reconstruct the field very similar to thermal simulation. Experimental reconstruction of both the temperature field and thermal deformation field by the reduced order model demonstrated its at least 360 times speedup with 99.8% accuracy compared to the actual thermal simulation. With the accurate reduced order model as a lightweight digital twin and using the differential evolution algorithm, three types of boundary conditions, i.e., the heat generation rates, the convective heat transfer coefficients and the thermal contact resistances, under the shaft rotation speed of 4,000r/min were identified within 11s. The maximum temperature simulation error was reduced from 85.6% to 6.6% and the thermal deformation simulation error was reduced from 60.9% to 10.8%.

Optimizing membrane reactor structures for enhanced hydrogen yield in CH4 tri-reforming: Insights from sensitivity analysis and machine learning approaches

This study explores the optimization of membrane reactor configurations to enhance hydrogen production through CH4 tri-reforming. The investigation employs ceramic membranes for oxygen, vapor, and carbon dioxide distribution within the reactor bed. A differential evolution algorithm is utilized alongside cuckoo search algorithm (CSA) and support vector regression (SVR) to determine optimal values for O2/CH4, H2O/CH4, and CO2/CH4 ratios, membrane thickness, and shell pressure, with hydrogen yield as the objective function. Results demonstrate that the oxygen membrane reactor achieves the highest hydrogen yield, reaching 2.02 and 1.75 for direct methanol synthesis and Fischer–Tropsch processes, respectively, representing a 7.98 % and 10.03 % increase compared to the conventional tri-reforming reactor. Furthermore, CSA and SVR emerge as invaluable tools, facilitating robust optimization and predictive modeling. The CSA efficiently navigates complex solution spaces to identify optimal parameters, while SVR accurately models relationships between input variables and hydrogen yield. Incorporating these methodologies enhances the effectiveness of membrane reactor design and synthesis gas production. This study contributes to advancements in clean energy technologies by providing insights into efficient hydrogen production methods using membrane reactors.

Reinforcement learning-based differential evolution algorithm for constrained multi-objective optimization problems

Many real-world problems can be established as Constrained Multi-objective Optimization Problems (CMOPs). It is still challenging to automatically set efficient parameters for Constrained Multi-Objective Evolutionary Algorithms (CMOEAs) to solve these CMOPs. A Reinforcement Learning-based Multi-Objective Differential Evolution (RLMODE) algorithm is proposed, in which the main parameters are dynamically adjusted. During the evolution process, the offspring generated is evaluated and compared with its corresponding parents, the relationship between the offspring and parent can adjust the parameters of RLMODE by the Reinforcement Learning (RL) technique. The feedback mechanism can produce the most appropriate parameters for RLMODE, which pushes the population towards feasible regions. The proposed RLMODE is evaluated on thirty functions and compared with some popular CMOEAs. The performance indicator IGD has revealed that the proposed RLMODE is competitive. Then, they are applied to solve the UAV path planning problem with three objectives and a constraint. The real application has further demonstrated the superiority of the proposed RLMODE.

Integrating Reinforcement Learning and Metaheuristics for Safe and Sustainable Health Tourist Trip Design Problem

This study aims to address the critical challenge of enhancing safety within the domain of health and wellness tourism through the framework of sustainable travel design. Given the rising demand for secure and sustainable tourism experiences, this investigation explores the integration of safety considerations into the strategic planning and execution of tourist journeys. The central focus of this research revolves around the development and application of the Modified Reinforcement Learning-Artificial Multiple Intelligence System (MRL-AMIS). Operating within a multi-objective framework, the objective of this study is to optimize essential parameters encompassing environmental sustainability, economic viability, tourist preferences, and societal acceptability, with a specific emphasis on safety. The efficacy of the proposed methodology is rigorously evaluated across various tourist group scenarios, providing a comprehensive assessment of its applicability. Our computational analyses unequivocally demonstrate the superiority of the proposed approach in enhancing safety, surpassing existing methods such as genetic algorithms, differential evolution algorithms, particle swarm optimization, etc., by 15%-25%. These findings underscore the effectiveness of the proposed method in seamlessly integrating safety considerations into the realm of sustainable tourism planning. This study highlights the potential of the proposed method as an advanced tool in the strategic planning of sustainable tourism, particularly within the context of health and wellness tourism. It constitutes a significant contribution to the field of sustainable tourism by offering an innovative approach that prioritizes safety while upholding the principles of environmental stewardship, economic viability, and societal sustainability.

Frequency Domain Channel-Wise Attack to CNN Classifiers in Motor Imagery Brain-Computer Interfaces.

Convolutional neural network (CNN), a classical structure in deep learning, has been commonly deployed in the motor imagery brain-computer interface (MIBCI). Many methods have been proposed to evaluate the vulnerability of such CNN models, primarily by attacking them using direct temporal perturbations. In this work, we propose a novel attacking approach based on perturbations in the frequency domain instead. For a given natural MI trial in the frequency domain, the proposed approach, called frequency domain channel-wise attack (FDCA), generates perturbations at each channel one after another to fool the CNN classifiers. The advances of this strategy are two-fold. First, instead of focusing on the temporal domain, perturbations are generated in the frequency domain where discriminative patterns can be extracted for motor imagery (MI) classification tasks. Second, the perturbing optimization is performed based on differential evolution algorithm in a black-box scenario where detailed model knowledge is not required. Experimental results demonstrate the effectiveness of the proposed FDCA which achieves a significantly higher success rate than the baselines and existing methods in attacking three major CNN classifiers on four public MI benchmarks. Perturbations generated in the frequency domain yield highly competitive results in attacking MIBCI deployed by CNN models even in a black-box setting, where the model information is well-protected. To our best knowledge, existing MIBCI attack approaches are all gradient-based methods and require details about the victim model, e.g., the parameters and objective function. We provide a more flexible strategy that does not require model details but still produces an effective attack outcome.

Fault Structure Reconstruction From Magnetic Anomalies Using an Improved Backtracking Search Optimization Algorithm

AbstractFault structure reconstruction is an important aspect of magnetic prospecting as it contributes to understanding tectonic history, hazard assessment, and infrastructure planning. Metaheuristics have proven useful in estimating fault structure parameters from geophysical anomalies. Recently, a modified version of the Backtracking Search Optimization Algorithm (BSA), named mBSA, has been proposed that combines mutation steps of BSA and Differential Evolution (DE) algorithms to achieve a better balance between exploration and exploitation. Here, we present imBSA as an efficient approach that builds on the improvements of mBSA. It uses an adaptive amplitude control factor in the mutation step derived from a normally distributed random number obtained via an exponential function. The performances of these approaches were tested on some widely used benchmark functions such as Rosenbrock, Rastrigin, Griewank, and Himmelblau. Synthetic magnetic fault anomalies were evaluated with and without noise, as well as three real anomalies from the Dehri (India), Perth Basin, and Lachlan Fold Belt (Australia). Prior to the optimizations, the solvability of the model parameters was examined using some error topography images. The obtained solutions in the synthetic cases were evaluated using principal component analysis. In addition, a novel visualization option for high‐dimensional optimization problems was introduced. The experiments showed that imBSA has a faster convergence rate and better optimization statistics compared to BSA and mBSA. Therefore, it is believed to be an efficient and good alternative tool to widely used metaheuristics such as Particle Swarm Optimization and DE in explaining geophysical anomalies and optimization problems in other geoscientific disciplines.

Research on Predicting Welding Deformation in Automated Laser Welding Processes with an Enhanced DEWOA-BP Algorithm

Research on Predicting Welding Deformation in Automated Laser Welding Processes with an Enhanced DEWOA-BP Algorithm

An adaptive differential evolution algorithm based on relaxation strategy to realize the Nash equilibrium of multi-oligopoly game

. Inspired by the idea of game learning, this paper studies the Cournot game model of the multi-oligopoly market and studies the realization of the Nash equilibrium by using adaptive differential evolution algorithm based on relaxation strategies. The realization path of Nash equilibrium is attained through an adaptive differential evolution based on a relaxation strategy procedure applied to an objective function with best response dynamics. Among them, the objective function is derived from the sum of the payoff changes of each oligopoly under best response decisions. And theoretically proved that the Nash equilibrium is the optimal solution of the objective function.

An industrial IoT-based deformation resistance prediction and thickness control method of cold-rolled strip in steel production systems

Different from traditional analytical and numerical simulation methods, data-driven methods can more effectively describe the effects of nonlinear and coupling factors. Although the fluctuations of deformation resistance and thickness for hot-rolled strips significantly influence the thickness accuracy in strip cold rolling, the traditional deformation resistance prediction model and thickness control mode don't involve that. Thus, based on the Industrial Internet of Things (IIOT) platform and data-driven technology, this work proposed a new method for deformation resistance prediction and a new mode of feed-forward control for strip thickness. IIOT platform can collect production data from the various levels to provide a complete data set. The deformation resistance prediction model is established by adopting the differential evolutionary algorithm to optimize the bi-directional long and short-term memory network. In addition, the feed-forward control (DFFC) strategy for strip thickness is proposed based on deformation resistance prediction. An application results of a 1420 mm production line of CR indicated that the established prediction model could provide a high accuracy prediction for deformation resistance, and compared to traditional thickness gauges-based automatic gauge control methods, the DDFC can capture the effect of the fluctuations of deformation resistance and thickness for the hot-rolled strip to improve thickness accuracy effectively.

Design and Comparative Analysis of E-Shape and H-Shape Microstrip Patch Antenna for IoT Application

In this study, H-type and E-type microstrip antenna arrays with high efficiency and management are designed using Surrogate Model Assisted Differential Evolution Algorithm (SADEA) for antenna synthesis. In order to meet the increasing wireless communication requirements of the internet of things (IoT) technology it has become very important to design new antenna designs with artificial intelligence techniques to transmit electromagnetic waves between antenna elements in the most efficient way. With a view to operate in the 2.5 GHz band, which is widely used in wireless communication technologies. H-type and E-type microstrip antennas with a 2x2 structure are analyzed and finally compared. By using the SADEA method, the amplitude and the distance between the array elements are optimized. The total amplitude of the 4-element E-type and H-type microstrip antenna array is minimized to 2.45885 V and 2.14929 V, respectively, after optimization, while it is 4V for both before optimization.

Design of information management model based on multiobjective optimization algorithm in intelligent electric financial system

The electric power infrastructure is the cornerstone of contemporary society’s sustenance and advancement. Within the intelligent electric power financial system, substantial inefficiency and waste in information management persist, leading to an escalating depletion of resources. Addressing diverse objectives encompassing economic, environmental, and societal concerns within the power system helps the study to undertake a comprehensive, integrated optimal design and operational scheduling based on a multiobjective optimization algorithm. This article centers on optimizing the power financial system by considering fuel cost, active network loss, and voltage quality as primary objectives. A mathematical model encapsulates these objectives, integrating equations and inequality constraints and subsequently introducing enhancements to the differential evolutionary algorithm. Adaptive variation and dynamic crossover factors within crossover, variation, and selection operations are integrated to optimize algorithm parameters, specifically catering to the multiobjective optimization of the electric power system. An adaptive grid method and cyclic crowding degree ensure population diversity and control the Pareto front distribution. They experimentally validated the approach and the comparisons conducted against AG-MOPSO, INSGA-II, and NSDE algorithms across standard test functions: ZDT1, ZDT2, ZDT3, and DTLZ4. The convergence evaluation indices for this study’s scheme on ZDT1 and ZDT2 are 0.000938 and 0.0034, respectively. Additionally, distribution evaluation indices on ZDT1, ZDT2, ZDT3, and ZDT4 stand at 0.0018, 0.0026, 0.0027, and 0.0009, respectively. These indices indicate a robust convergence and distribution, facilitating the optimization of electric power financial information management and the intelligent handling of the electric power financial system’s information, thereby enhancing the allocation of material and financial resources.