Evolutionary Algorithm Research Articles

Neural network controller for hybrid energy management system applied to electric vehicles

Hybrid energy storage systems based on batteries and supercapacitors can mitigate the aging of electric vehicle batteries aging by avoiding high currents and rapid discharge cycles. This system requires energy management systems that efficiently split the power during real driving cycles. This paper outlines a design methodology for creating a Multilayer Perceptron neural controller that governs the power distribution between the storage system components. In parallel, a Proportional-Integral-Derivative controller was implemented to ensure voltage regulation in the primary DC bus, ensuring stable operation of the entire system and providing flexibility to the neural design. The controller was tuned using particle swarm optimization, hippopotamus optimization, and differential evolution algorithms, designed to minimize the battery root-mean-square current in a comprehensive vehicle simulation. The training was structured into layers to approach the physical problem and controller optimization facets. The controller’s primary goal is to minimize the battery strain, mitigating stress events and prolonging its lifespan. The energy management neural controller was designed using a simple drive cycle and later validated with a realistic cycle. The findings demonstrate the feasibility of achieving a significant reduction in current, both in peak value and on average, especially when compared to basic energy management strategies. The process uncovered a strategy that prioritizes the use of the supercapacitor in power balance, with this effect being more pronounced during critical load events. The main bus voltage remained stable, with no deviations exceeding 5%, owing to the voltage regulator’s stability. Additionally, the particle swarm optimization and the hippopotamus optimization techniques exhibited notable performance compared to the differential evolution algorithm for this problem. Subsequently, the design was evaluated in a more complex cycle, revealing a slight decrease in average battery current performance. However, it still maintained a significant reduction in peak battery current compared to traditional controllers. Nevertheless, this methodology demonstrated power management optimization efficacy and can be extended to more complex scenarios.

Design of Adaptive Fuzzy Modeling Based on Evolutionary Optimization Algorithms: A Comparative Study

Design of Adaptive Fuzzy Modeling Based on Evolutionary Optimization Algorithms: A Comparative Study

Short-term Traffic Flow Prediction：A Method of MEA-LSTM Model Based on Chaotic Characteristics Analysis

To effectively improve the accuracy of short-term traffic flow prediction, an improved Long Short-Term Memory (LSTM) method is proposed using the Mind Evolution Algorithm (MEA). Firstly, to address the issues of abnormal and missing traffic flow data, a Neighborhood Stacked Denoising AutoEncoder (NSDAE) is used for data repair. Then, the maximum Lyapunov exponent is used to determine the chaotic characteristics. Meanwhile, based on Bayesian estimation theory, the features of three-parameter sequences are fused in high-dimensional space using phase space reconstruction technique to obtain reconstructed multi-parameter fused traffic flow data. Finally, by taking advantage of the fact that MEA can divide the data into several subpopulations for optimal search separately, a prediction model based on MEA to improve LSTM is proposed. The results show that compared to the other two traditional data restoration methods, the NSDAE has higher accuracy, with the lowest average values of RMSE, MAE, and MAPE. Through the phase space reconstruction technique, the feature fusion of three parameters of traffic flow is realized in high-dimensional space, which makes up for the insufficiency of a single time-series data that cannot comprehensively levy the characteristics of traffic flow. The MEA-LSTM model outperforms the LSTM model in terms of prediction accuracy, computational efficiency, and generalization ability, and its RMSE, MEA, and MAPE are reduced by 24.3%, 28.9%, and 30.1%, respectively.

Production Quality Evaluation of Electronic Control Modules Based on Deep Belief Network

The electronic control module is an important part of a digital electronic detonator, which undergoes a complex production process that includes three electrical performance tests and three visual inspection procedures. In each inspection procedure, several different types of data are generated daily, including numerical and categorical data. To evaluate the production quality of electronic control modules, an algorithm based on a Deep Belief Network with Multi-mutation Differential Evolution (MDE-DBN) is designed in this study. First, key indicators are extracted to construct a production quality evaluation index system. A Multi-mutation Differential Evolution algorithm is designed to optimize the initial network weights of the Deep Belief Network (DBN) and integrate the production quality information into the pre-training phase. Subsequently, the preprocessed experimental data are input into the MDE-DBN algorithm to obtain the distributions of excellent, general, and unqualified production statuses, verifying the effectiveness of the algorithm. The experimental results show that the MDE-DBN algorithm has significant advantages in evaluation accuracy when compared with DBNs improved by other intelligent optimization algorithms and machine learning methods.

Evolutionary multitasking algorithm based on a dynamic solution encoding strategy for the minimum s-club cover problem

Evolutionary multitasking algorithm based on a dynamic solution encoding strategy for the minimum s-club cover problem

Two-stage bidirectional coevolutionary algorithm for constrained multi-objective optimization

Objective optimization and constraint satisfaction are two primary and conflicting tasks in solving constrained multi-objective optimization problems (CMOPs). To better trade off them, this paper proposes a two-stage bidirectional coevolutionary algorithm, termed C-TBCEA, for constrained multi-objective optimization. It consists of two stages, with each concentrating on specific targets, i.e., the first stage primarily focuses on objective optimization while the second stage focuses on constraint satisfaction by employing different evolutionary strategies at each stage. Via the synergy of the two stages, a dynamic trade-off between objective optimization and constraint satisfaction can be achieved, thus overcoming the distinctive challenges that may be encountered at different stages of evolution. In addition, to take advantage of both feasible and infeasible solutions, we employ two populations, i.e., the main population that stores the non-dominated feasible solutions and the archive population that maintains the informative infeasible solutions, to prompt the bidirectional coevolution of them. To validate the effectiveness of the proposed C-TBCEA, experiments are carried out on 6 CMOP test suites and 17 real-world CMOPs. The results demonstrate that the proposed algorithm is very competitive with 9 state-of-the-art constrained multi-objective optimization evolutionary algorithms (CMOEAs).

An improved reinforcement learning-based differential evolution algorithm for combined economic and emission dispatch problems

To overcome challenges posed by escalating environmental pollution and climate change, the combined economic and emission dispatch problem is proposed to balance economic efficiency with emission cost. The primary objective of the problem is to ensure that emissions are minimized while optimal economic costs are achieved simultaneously. However, due to the nonlinear and nonconvex characteristics of the model, the optimization is confronted with many difficulties. Hence, an innovative improved reinforcement learning-based differential evolution algorithm is proposed in this article, with reinforcement learning seamlessly integrated into the differential evolution algorithm. Q-learning from reinforcement learning technique is utilized to dynamically adjust parameter settings and select appropriate mutation strategies, thereby boosting the algorithm's adaptability and overall performance. The effectiveness of the proposed algorithm is tested on thirty testing functions and combined economic and emission dispatch problems in comparison with the other five algorithms. According to the experimental results of testing functions, superior performance is consistently achieved by the proposed algorithm, with the highest adaptability exhibited and an average ranking of 1.4167. Its superiority is further demonstrated through Wilcoxon tests on results of testing functions and combined economic and emission dispatch problems with the proportion of 100%, and the proposed algorithm is significantly better than other algorithms at a 0.05 significance level. The superiority of the proposed algorithm in optimizing combined economic and emission dispatch problems demonstrates that the proposed algorithm is shown to be adaptable to complex optimization environments, which proves useful for industrial applications and artificial intelligence.

Precision parameter estimation in Proton Exchange Membrane Fuel Cells using depth information enhanced Differential Evolution

Proton Exchange Membrane Fuel Cell (PEMFC) models require parameter tuning for their design and performance improvement. In this study, Depth Information-Based Differential Evolution (Di-DE) algorithm, a novel and efficient metaheuristic approach, is applied to the complex, nonlinear optimization problem of PEMFC parameter estimation. The Di-DE algorithm was tested on twelve PEMFCs (BCS 500 W PEMFC, Nedstack 600 W PS6 PEMFC, SR-12 500 W PEMFC, H-12 PEMFC, STD 250 W PEMFC, HORIZON 500 W PEMFC and four 250W PEMFC and two H-12 12W PEMFC) and showed excellent accuracy. The Di-DE algorithm is was compared with other advanced evolutionary algorithms like iwPSO, CLPSO, DNLPSO, SLPSO, SaDE, SHADE, JADE, QUATRE, LSA, QUATRE-EMS and C-QUATRE, which obtained a minimum objective function value of 0.0255 and an average runtime improvement of 98.8%. The optimized parameters of the proposed method yielded the Sum of Squared Errors (SSE) as low as 0.00002 in some cases, which indicates better precision and stability. Moreover, the voltage–current (V–I) and power–voltage (P–V) characteristics predicted by Di-DE were within 1% error relative to the experimental data for all tested PEMFCs. The results of this work highlight the potential of the Di-DE algorithm to enable more sophisticated modelling and optimization of PEMFCs, which in turn will help to broaden the use of PEMFCs in clean energy applications.

A comparison of different clustering algorithms for the project time buffering problem

This paper studies the decentralised time buffering problem (TBP) to absorb project risk by building sufficient buffers with the aim of obtaining a stable project schedule. First, the position of the buffers in the project network should be determined and, subsequently, each buffer must be optimally sized. We investigate different activity clustering methods (K-means, rank order, criticality-based and network clustering) to determine the ideal groups of activities to be clustered together and protected by an allocated buffer. The obtained clusters of activities are then inputted in a multi-population multi-factorial evolutionary algorithm (MPMFEA) for creating buffers based on the characteristics of the activities in each cluster. To the best of our knowledge, this is the first study to integrate existing clustering methods into a buffering algorithm in order to optimise the project stability. Previous studies hybridising both methods use a single clustering algorithm (e.g. K-means) that does not use the same information than the buffering algorithm or require more complex (simulation-based) buffering methods. The computational experiments on a large set of artificial instances validate the effectiveness of the proposed MPMFEA for solving the TBP, especially in combination with the network clustering method. Although the generic K-Means method is still considered a viable option for clustering, the more pragmatic clustering methods are more effective. We inform project managers that considering precedence relations between activities during clustering is crucial, but mimicking this behaviour in all clustering methods does not guarantee successful protection of their projects.

A comprehensive multi-stage decision-making model for supplier selection and order allocation approach in the digital economy

The increasingly serious environmental issues and fierce competition caused by globalization have brought pressure on supply chain managers who seek to allocate multiple purchase demands comprehensively, highlighting the significance of supplier assessment considering sustainability and technique. Moreover, many multi-criteria decision-making (MCDM) methods fail to quantify the risk preference of decision-makers (DMs) when conducting the supplier assessment process. Indeed, a hybrid supplier selection and order allocation model that integrates such requirements is yet to be proposed. Thus, this work develops a comprehensive decision-making model that constructs a deep learning model to forecast the potential demand and addresses the sustainable supplier selection based on cumulative prospect theory (CPT) and multi-material order allocation problem simultaneously. The proposed order allocation model is solved by the second generation of adaptive geometry estimation based many-objective evolutionary algorithm, with the technique of order preference similarity to the ideal solution used to filter out the best Pareto solution for DMs as the reference. Through implementing an illustrative case study of a leading Chinese engineering machinery manufacturer followed by a sensitivity analysis, the relatively strong applicability and scalability of the proposed model and methods are demonstrated. The results show that introducing Weibull distribution to estimate the theoretical obsolescence rate of historically sold accessories can result in higher demand prediction accuracy for consumable mechanical accessories. Integrating CPT into the MCDM framework allows us to evaluate suppliers more comprehensively by capturing the effect of DMs’ risk preferences and gain or loss sensitivity.

Obtaining diverse solutions for reliable multi‐microgrid network design

AbstractThe multi‐microgrid network structure optimisation problem (MNSDOP) aims to minimise the circuit lengths while maximising power transmission stability, a crucial aspect for system robustness enhancement. While existing literature primarily focuses on achieving Pareto optimal solutions there is an overlooked emphasis on solution diversity. In situations like MNSDOPs, several optimal solutions might possess identical or comparable objective values, classifying them under multimodal multi‐objective optimisation roblems (MMOPs). Offering an array of optimal solutions equips decision‐makers with a holistic understanding of the problem, aiding in the identification of preferred solutions. Motivated by this gap, we introduce a multimodal multi‐objective evolutionary algorithm tailored for MNSDOPs, termed MMO‐BM. We also propose a diversity evaluation metric specifically for binary optimisation problems, substantially amplifying the exploration capabilities of evolutionary algorithms. In addition, a novel matrix‐oriented DE operator is proposed to accerate the convergence process. Experimental evaluations attest to our method's prowess in yielding diverse and high‐quality solutions.

Bayesian Inverse Transfer in Evolutionary Multiobjective Optimization

Transfer optimization enables data-efficient optimization of a target task by leveraging experiential priors from related source tasks. This is especially useful in multiobjective optimization settings where a set of tradeoff solutions is sought under tight evaluation budgets. In this article, we introduce a novel concept of inverse transfer in multiobjective optimization. Inverse transfer stands out by employing Bayesian inverse Gaussian process models to map performance vectors in the objective space to population search distributions in task-specific decision space, facilitating knowledge transfer through objective space unification . Building upon this idea, we introduce the first Inverse Transfer Evolutionary Multiobjective Optimizer (invTrEMO). A key highlight of invTrEMO is its ability to harness the common objective functions prevalent in many application areas, even when decision spaces do not precisely align between tasks. This allows invTrEMO to uniquely and effectively utilize information from heterogeneous source tasks as well. Furthermore, invTrEMO yields high-precision inverse models as a significant byproduct, enabling the generation of tailored solutions on-demand based on user preferences. Empirical studies on multi- and many-objective benchmark problems, as well as a practical case study, showcase the faster convergence rate and modeling accuracy of the invTrEMO relative to state-of-the-art evolutionary and Bayesian optimization algorithms. The source code of the invTrEMO is made available at https://github.com/LiuJ-2023/invTrEMO .

Intelligent decision-making approach for rapid optimization of double-wall cooling structures under varying cooling demands

Double-wall structures are widely employed for cooling turbine blades. When manufacturing materials and application conditions vary, the design of double-wall structures should be adapted to meet different cooling demands. However, because of inevitable repetitive iterations, conventional evolutionary algorithms exhibit low efficiency in optimizing double-wall structures when cooling demands change. This paper presents a reinforcement learning-based method to optimize double-wall cooling structures facing various cooling demands rapidly. A decision network established a forward mapping from cooling demands to optimized structural designs and was trained using historical decision-making experiences generated from numerical simulations. The historical experiences were quantified uniformly based on the cooling performance gains or losses resulting from changes in geometric parameters. The physical understanding that cooling performance evaluations differ with varying cooling demands was incorporated into the training dataset. The performance of the trained decision network was validated by varying the cooling optimization objectives. Results indicate that the decision network can achieve optimal double-wall cooling units that meet given objectives with a single decision. The decision process took 5 ms, nearly 90 times faster than traditional genetic algorithms. Moreover, a partitioning, multi-cycle application strategy for the decision network was employed to optimize double-wall cooling plates under non-uniform inlet temperature distributions, resulting in a 13.7K reduction in overall average temperature and improved radial temperature uniformity.

Online public opinion prediction based on a novel conformable fractional discrete grey model

PurposeAs social networks have developed to be a ubiquitous platform of public opinion spreading, it becomes more and more crucial for maintaining social security and stability by accurately predicting various trends of public opinion dissemination in social networks. Considering the fact that the dissemination of online public opinion is a dynamic process full of uncertainty and complexity, this study establishes a novel conformable fractional discrete grey model with linear time-varying parameters, namely the CFTDGM(1,1) model, for more accurate prediction of online public opinion trends.Design/methodology/approachFirst, the conformable fractional accumulation and difference operators are employed to build the CFTDGM(1,1) model for enhancing the traditional integer-order discrete grey model with linear time-varying parameters. Then, to improve forecasting accuracy, a base value correction term is introduced to optimize the iterative base value of the CFTDGM(1,1) model. Next, the differential evolution algorithm is selected to determine the optimal order of the proposed model through a comparison with the whale optimization algorithm and the particle swarm optimization algorithm. The least squares method is utilized to estimate the parameter values of the CFTDGM(1,1) model. In addition, the effectiveness of the CFTDGM(1,1) model is tested through a public opinion event about “IG team winning the championship”. Finally, we conduct empirical analysis on two hot online public opinion events regarding “Chengdu toddler mauled by Rottweiler” and “Mayday band suspected of lip-syncing,” to further assess the prediction ability and applicability of the CFTDGM(1,1) model by comparison with seven other existing grey models.FindingsThe test case and empirical analysis on two recent hot events reveal that the CFTDGM(1,1) model outperforms most of the existing grey models in terms of prediction performance. Therefore, the CFTDGM(1,1) model is chosen to forecast the development trends of these two hot events. The prediction results indicate that public attention to both events will decline slowly over the next three days.Originality/valueA conformable fractional discrete grey model is proposed with the help of conformable fractional operators and a base value correction term to improve the traditional discrete grey model. The test case and empirical analysis on two recent hot events indicate that this novel model has higher accuracy and feasibility in online public opinion trend prediction.

AI-Based Supply Chain Risk Management

As the complexity of global supply chains increases, supply chain risk management becomes increasingly important. The development of artificial intelligence (AI) technology provides new tools and methods for managing supply chain risks. This paper aims to review the application of AI in supply chain risk management, exploring current research progress, challenges, and future development trends. Through a systematic literature search and analysis, this paper examines relevant papers from databases such as Google Scholar, Web of Science, EI, and Scopus, discussing the specific applications of machine learning, deep learning, neural networks, fuzzy logic, genetic algorithms, and evolutionary algorithms in supply chain risk management. The study finds that these AI technologies demonstrate significant advantages in risk prediction, anomaly detection, image recognition, text mining, supply chain optimization, and emergency strategy formulation. However, issues related to data privacy and security, technical complexity, and implementation difficulties remain major challenges in current applications. In the future, as AI technology advances and interdisciplinary integration develops, supply chain risk management will have more opportunities. This study provides specific application suggestions for enterprises and decision-makers and points the way for future research.

Multi‐Objective Evolutionary Algorithm Based on Decomposition With Orthogonal Experimental Design

ABSTRACTMulti‐objective evolutionary optimisation algorithms (MOEAs) have become a widely adopted way of solving the multi‐objective optimisation problems (MOPs). The decomposition‐based MOEAs demonstrate a promising performance for solving regular MOPs. However, when handling the irregular MOPs, the decomposition‐based MOEAs cannot offer a convincing performance because no intersection between weight vector and the Pareto Front (PF) may lead to the same optimal solution assigned to the different weight vectors. To solve this problem, this paper proposes an MOEA based on decomposition with the orthogonal experimental design (MOEA/D‐OED) that involves the selection operation, Orthogonal Experimental Design (OED) operation, and adjustment operation. The selection operation is to judge the unpromising weight vectors based on the history data of relative reduction values and convergence degree. The OED method based on the relative reduction function could make an explicit guidance for removing the worthless weight vectors. The adjustment operation brings in an estimation indicator of both diversity and convergence for adding new weight vectors into the interesting regions. To verify the versatility of the proposed MOEA/D‐OED, 26 test problems with various PFs are evaluated in this paper. Empirical results have demonstrated that the proposed MOEA/D‐OED outperforms eight representative MOEAs on MOPs with various types of PFs, showing promising versatility. The proposed algorithm shows highly competitive performance on all the various MOPs.

Reliable and cost-efficient session provisioning in CRNs using spectrum sensing as a service

With the advancement of wireless communication technologies, and the growing number of wireless and IoT applications that demand various types and volumes of data, Sensing as a Service (SaaS) has emerged as a necessary enabling business model for many of those applications. Spectrum Sensing as a Service (SSaaS) has also emerged as a form of SaaS that is concerned with the monitoring of wireless spectrum to facilitate its safe reuse by cognitive radio-enabled wireless users. SSaaS was primarily motivated by the need for a low-cost, accurate, and reliable spectrum sensing service to support a plethora of heterogeneous wireless devices and applications. Under the SSaaS model, clients need to pay the service provider for the sensing service they receive. In this paper, we address the problem of allocating spectrum channels to links of a given communication session in a cognitive radio network (CRN) that utilizes SSaaS. The objective is to allocate channels such that the worst link availability among the session is maximized and the spectrum access cost is minimized. A number of multi-objective evolutionary optimization algorithms (MOEAs) were used to solve this multi-objective optimization problem. Extensive experimentation was conducted to compare between these algorithms and identify the best ones to use. We also propose a post-processing greedy algorithm to further enhance the solution obtained by a MOEA algorithm. Results show that an improvement of up to 20% can be achieved using the proposed greedy algorithm under some network settings.

Solving coupled non-linear schrödinger equations via quantum imaginary time evolution

AbstractCoupled non-linear Schrödinger equations are crucial in describing dynamics of many-particle systems. We present a quantum imaginary time evolution (ITE) algorithm as a solution to such equations in the case of nuclear Hartree-Fock approach. Under a simplified Skyrme interaction model, we calculate the ground state energy of an oxygen-16 nucleus and demonstrate that the result is in agreement with the classical ITE algorithm. We examine bottlenecks and deficiencies in the quantum algorithm and suggest possible improvements.

A sustainable vaccine supply-production-distribution network with heterologous and homologous vaccination strategies: Bi-objective optimization

Heterologous and homologous Coronavirus Disease 2019 (COVID-19) vaccination against Severe Acute Respiratory Syndrome (SARS)-CoV-2 are robust and proactively adaptable strategies. However, there is still a lack of appropriate mathematical models for integrating vaccination strategies into the vaccine supply chain network. This study develops a supply-production-distribution-inventory-allocation problem in the Sustainable Vaccine Supply-Production-Distribution Network (SVSPDN) to fill this gap for the first time. The outstanding novelties of this research are prioritizing vaccines and sequencing injection doses to increase vaccination effectiveness. In addition, the remarkable new contribution of the proposed mathematical model is the design of new bi-objective, multi-dose, multi-level, and multi-period to ensure the sustainability performance of the entire network. This aim is achievable by minimizing the cost of supplying, producing, and distributing vaccines and fulfilling social goals by maximizing vaccination effectiveness. Also, a scenario-based robust stochastic optimization approach is presented to handle uncertainties. Since the SVSPDN design is an NP-hard problem, to solve the proposed mathematical model, three Pareto-based evolutionary algorithms, including Non-Dominated Sorting Genetic Algorithm II (NSGA-II), Multi-Objective Particle Swarm Optimization (MOPSO), and Multi-Objective Gray Wolf Optimizer (MOGWO), are applied. The Taguchi design method is applied to tuning the parameters due to the sensitivity of meta-heuristic algorithms to input parameters. Then, a comparison is performed using four assessment metrics, including the Number of Pareto Solutions (NPS), Diversification Matrix (DM), Mean Ideal Distance (MID), Spread of Non-Dominance Solutions (SNS), and Computation Time (CT). The results reveal that the NSGA-II and MOGWO algorithms have performances that are very close to each other. However, MOGWO performs better in tackling the problem and is superior to the NSGA-II and MOPSO regarding assessment metrics and computation time. A case study of Iran is investigated to indicate the efficiency and applicability of the proposed model. Finally, sensitivity analyses, managerial insights, and practical implications are discussed.

A Q-learning driven multi-objective evolutionary algorithm for worker fatigue dual-resource-constrained distributed hybrid flow shop

In practical industrial production, workers are often critical resources in manufacturing systems. However, few studies have considered the level of worker fatigue when assigning resources and arranging tasks, which has a negative impact on productivity. To fill this gap, the distributed hybrid flow shop scheduling problem with dual-resource constraints considering worker fatigue (DHFSPW) is introduced in this study. Due to the complexity and diversity of distributed manufacturing and multi-objective, a Q-learning driven multi-objective evolutionary algorithm (QMOEA) is proposed to optimize both the makespan and total energy consumption of the DHFSPW at the same time. In QMOEA, solutions are represented by a four-dimensional vector, and a decoding heuristic that accounts for real-time worker productivity is proposed. Additionally, three problem-specific initialization heuristics are developed to enhance convergence and diversity capabilities. Moreover, encoding-based crossover, mirror crossover and balanced mutation methods are presented to improve the algorithm’s exploitation capabilities. Furthermore, a Q-learning based local search is employed to explore promising nondominated solutions across different dimensions. Finally, the QMOEA is assessed using a set of randomly generated instances, and a detailed comparison with state-of-the-art algorithms is performed to demonstrate its efficiency and robustness.