Differential Evolution Strategy Research Articles

Stochastic configuration networks for short-term power load forecasting

The accuracy of power load forecasting can enhance the operational efficacy and economic efficiency of the power system. This paper utilizes the stochastic configuration network (SCN) and improved marine predator algorithm (IMPA) for power systems short-term load forecasting. First, the daily load data is clustered using the kernel fuzzy C-means (KFCM), and the clustering results are utilized as training input data for the load forecasting model. Then, load forecasting is performed using SCN. Next, IMPA is developed by integrating the opposite-based differential evolution (ODE) and normal mutation perturbation strategy for the purpose of preventing the algorithm from converging to local minima. Finally, IMPA is employed to optimize the critical parameters of the load forecasting model. The experimental results clearly demonstrate that the proposed algorithm performs favorably in terms of the prediction accuracy compared to other modeling algorithms.

Optimizing efficiency of proton exchange membrane electrolyzer system based on multiphysics model and differential evolution strategy

This work focuses on a comprehensive performance enhancement considering hydrogen production, energy consumption of balance of plant (BOP), and thermal safety by optimizing the inlet water temperature and flow velocity. First, a multiphysics model is introduced to show that the inlet water temperature and flow velocity can affect the current density as well as the hydrogen production. An overall efficiency is defined to balance the hydrogen production and the energy consumption of BOP. After that, a surrogate model is established based on artificial neural network to replace the multiphysics model because of its huge computational effort. For the thermal safety and performance enhancement, an optimization problem with the maximum membrane temperature as a constraint and the overall efficiency as an objective function is proposed. The differential evolution (DE) strategy with the feasibility rule (FR) is adopted to solve this optimization problem. Finally, the DE-FR-based optimization algorithm is applied to three typical days in Chengdu of China representing summer, winter, and transitional seasons, respectively. Compared to the invariant inlet water flow velocity and temperature, the optimal and time-varying parameters improve their efficiency by relative percentages of 2.68%, 3.79%, and 3.29% for the three typical days, respectively.

Enhanced Prairie Dog Optimization with Differential Evolution for solving engineering design problems and network intrusion detection system

This paper introduces a novel hybrid optimization algorithm, PDO-DE, which integrates the Prairie Dog Optimization (PDO) algorithm with the Differential Evolution (DE) strategy. This research aims to develop an algorithm that efficiently addresses complex optimization problems in engineering design and network intrusion detection systems. Our method enhances the PDO's search capabilities by incorporating the DE's principal mechanisms of mutation and crossover, facilitating improved solution exploration and exploitation. We evaluate the effectiveness of the PDO-DE algorithm through rigorous testing on 23 classical benchmark functions, five engineering design problems, and a network intrusion detection system (NIDS). The results indicate that PDO-DE outperforms several state-of-the-art optimization algorithms regarding convergence speed and accuracy, demonstrating its robustness and adaptability across different problem domains. The PDO-DE algorithm's potential applications extend to engineering challenges and cybersecurity issues, where efficient and reliable solutions are critical; for example, the NIDS results show significant results in detection rate, false alarm, and accuracy with 98.1%, 2.4%, and 96%, respectively. The innovative integration of PDO and DE contributes significantly to stochastic optimization and swarm intelligence, offering a promising new tool for tackling diverse optimization problems. In conclusion, the PDO-DE algorithm represents a significant scientific advancement in hybrid optimization techniques, providing a more effective approach for solving real-world problems that require high precision and optimal resource utilization.

Multi-AUVs cooperative path planning in 3D underwater terrain and vortex environments based on improved multi-objective particle swarm optimization algorithm

This paper presents an improved multi-objective particle swarm optimization (IMOPSO) algorithm for cooperative path planning of multiple autonomous underwater vehicles (multi-AUVs) in 3D underwater terrain and vortex environment. Traditional PSO algorithms often struggle with local optima in high-dimensional, complex search spaces. To address this, we propose enhancements including Sine double-sequence perturbation interpolation for initial particle population generation, adaptive differential evolution strategies, iterative local search, and traction operations to enhance global search capability and avoid local optima. Additionally, cubic spline interpolation is employed to smooth the planned paths. The objective function is tailored to the ocean currents environment, considering path length, smoothness, deflection angle, and current impact, with a penalty function method to handle multi-AUVs cooperation and environmental constraints optimization for optimal energy consumption. Simulation results demonstrate that the proposed algorithm effectively plans short-range, high-security global paths for multi-AUVs, verifying its effectiveness and practicality in meeting the navigation requirements in ocean currents environments.

Nondestructive and accurate prediction of IGBT junction temperature in wind power converters: based on IDMOA-ELM prediction method

ABSTRACT This study has made significant advancements in predicting junction temperature of IGBTs in wind power converters. An improved dwarf mongoose optimisation algorithm (IDMOA) with superior performance was developed. The IDMOA improves the initial solution quality through Hammersley initialisation, balances global search and local optimisation with nonlinear dynamic decreasing weight factors, and enhances population diversity via differential evolution strategy to avoid local optima. The IDMOA optimizes random parameters of extreme learning machine (ELM), overcoming the limitations of traditional manual parameter setting and constructing a high-precision IDMOA optimised ELM (IDMOA-ELM) model. Simulation results indicate that the IDMOA-ELM excels in predicting the IGBT junction temperature, achieving a low RMSE value of 0.1034 ℃. This significantly surpasses existing technologies and provides robust support for temperature management in wind power converters. Furthermore, to verify the performance of the DMOA-ELM in practical application scenarios, IGBT accelerated ageing experiments were conducted, yielding a comprehensive validation dataset. The high accuracy and effectiveness of the IDMOA-ELM were confirmed through comparative analysis with experimental data, enhancing the credibility of the research. This study achieved non-destructive and high-precision prediction of IGBT junction temperature in wind power converters and significantly improving the reliability of wind power converters.

Neuroevolution reinforcement learning for multi-echelon inventory optimization with delivery options and uncertain discount

The advanced information technology has enabled supply chain to make centralized optimal decision, allowing to make a global optimal solution. However, dealing with uncertainty is important in inventory management. Besides demand and supply uncertainties, supplier discounts also often arise unexpectedly. Further, suppliers or third-parties typically offer various delivery options in which trade-off occurs between cost and lead time. Thus, this study introduces new problem namely Multi-Echelon Inventory Optimization with Delivery Options and Uncertain Discount (MEIO-DO-UD). As a solution, Neuroevolution Reinforcement Learning (NERL) framework is developed for minimizing total system cost. The environment is modeled via System Dynamics (SD) and actor is presented by integration of Artificial Neural Network and Evolutionary Algorithm (EA), creating an effective decision-making model under dynamic uncertainty. The experimental study has been conducted where two different supply chain networks are given namely serial and divergence. Three EA algorithms are compared namely Differential Evolution (DE), Memetic Algorithm (MA), and Evolution Strategy (ES). Furthermore, NERL is also compared with the EA-optimized classical continuous review model namely (s,Q). The result shows that regardless what EA type is used, the proposed NERL always outperforms EA-optimized (s,Q) model. The more complex the problem, the further improvement can be made i.e. cost reduction up to 58%, followed by the fill rate improvement. The result also shows that NERL can avoid overfitting. Managerial implications are highlighted where NERL provides the more stable inventory level among all supply chain partners and bull-whip effect can be damped.

Reentry trajectory optimization of hypersonic glide vehicle based on improved particle swarm algorithm

The reentry process of hypersonic glide vehicle presents a host of intricate constraint issues. A method for optimizing the reentry trajectory of hypersonic glide vehicle is presented to achieve the shortest possible reentry time. The approach utilizes the improved particle swarm optimization (IPSO) algorithm. Firstly, population initialization is performed using the quasi-oppositional differential evolution (QODE) strategy to promote diversity within the population. Then an adaptive adjustment algorithm is designed for the inertia weight coefficient and learning factor to achieve a balance between global and local search capabilities. Finally, utilizing the IPSO algorithm for optimizing the flight parameters of the reentry trajectory, the reentry time is reduced from 338 seconds to 335 seconds. The simulations results indicate that the IPSO algorithm outperforms both the standard PSO algorithm and genetic algorithm (GA) in terms of optimization performance index.

Distribution Network Reconfiguration Based on an Improved Arithmetic Optimization Algorithm

Aiming to address the defects of the arithmetic optimization algorithm (AOA), such as easy fall into local optimums and slow convergence speed during the search process, an improved arithmetic optimization algorithm (IAOA) is proposed and applied to the study of distribution network reconfiguration. Firstly, a reconfiguration model is established to reduce network loss, and a cosine control factor is introduced to reconfigure the math optimization accelerated (MOA) function to coordinate the algorithm’s global exploration and local exploitation capabilities. Subsequently, a reverse differential evolution strategy is introduced to improve the overall diversity of the population and Weibull mutation is performed on the better-adapted individuals generated in each iteration to ensure the quality of the optimal individuals generated in each iteration and strengthen the algorithm’s ability to approach the optimal solution. The performance of the improved algorithm is also tested using eight basis functions. Finally, simulation analysis is carried out by taking the IEEE33 and IEEE69 node systems and a real power distribution system as examples; the results show that the proposed algorithm can help to reconfigure the system quickly, and the system node voltages and network losses were significantly improved after the reconfiguration.

Evolutionary multitasking for multiobjective optimization based on hybrid differential evolution and multiple search strategy

Multitask optimization uses the knowledge transfer between tasks to deal with multiple related tasks simultaneously, which obtains better optimization performance. Currently, most multitask optimization algorithms have shortings such as slow convergence speed, difficulty in generating high-quality solutions and insufficient population diversity. To solve these problems, an new multiobjective multitask evolutionary algorithm is proposed in this paper. The algorithm introduces hybrid differential evolution strategy and multiple search strategy to generate offspring and high-quality solutions. In order to accelerate the population convergence and improve the lack of population diversity, we mixed two differential mutation operators with different functions to generate offspring. While generating high-quality solutions to improve the population convergence speed, some random solutions can also be generated to maintain the population diversity, which enhances the ability of the algorithm to jump out of the local optimal. Secondly, the multiple search strategy uses the information collected from multiple dimensions to optimize decision variables, and generates high-quality solutions by migrating useful knowledge. Finally, the performance of the algorithm is verified on two classical multitask test set. Compared with other related algorithms, the proposed algorithm has a faster convergence speed and better distribution performance, and shows a good ability to jump out of local optimal.

MSAO: A multi-strategy boosted snow ablation optimizer for global optimization and real-world engineering applications

Snow Ablation Optimizer (SAO) is a cutting-edge nature-inspired meta-heuristic technique that mimics the sublimation and melting processes of snow in its quest for optimal solution to complex problems. While SAO has demonstrated competitive performance in comparison to classical algorithms in early research, it still exhibits certain limitations including low convergence accuracy, a lack of population diversity, and premature convergence, particularly when addressing high-dimensional intricate challenges. To mitigate the above-mentioned adverse factors, this paper introduces a novel variant of SAO with featuring four enhancement strategies collectively referred as MSAO. Firstly, the good point set initialization strategy is employed to generate a uniformly distributed high-quality population, which facilitates the algorithm to enter the appropriate search domain rapidly. Secondly, the greedy selection method is adopted to reserve better candidate solutions for the next iteration, thus striking a robust exploration–exploitation balance. Then, the Differential Evolution (DE) scheme is introduced to expand the search range and enhance the exploitation capability of the algorithm for higher convergence accuracy. Finally, to reduce the risk of falling into local optima, a Dynamic Lens Opposition-Based Learning (DLOBL) strategy is developed to operate on the current optimal solution dimension by dimension. With the blessing of these strategies, the optimization performance of MSAO is comprehensively improved. To comprehensively evaluate the optimization performance of MSAO, a series of numerical optimization experiments are conducted using the IEEE CEC2017 & CEC2022 test sets. In the IEEE CEC2017 experiments, the optimal crossover probability CR=0.8 is determined and the effectiveness of each improvement strategy is ablatively verified. MSAO is compared with the basic SAO, various state-of-the-art optimizers, and CEC2017 champion algorithms in terms of solution accuracy, convergence speed, robustness, and scalability. In the IEEE CEC2022 experiments, MSAO is compared with some recently developed improved algorithms to further validate its superiority. The results demonstrate that MSAO has excellent overall optimization performance, with the smallest Friedman mean rankings of 1.66 and 1.25 on both test suites, respectively. In the majority of test cases, MSAO can provide more accurate and reliable solutions than other competitors. Furthermore, six realistic constrained engineering design challenges and one photovoltaic model parameter estimation issue are employed to demonstrate the practicality of MSAO. Our findings suggest that MSAO has excellent optimization capacity and broad application potential.

Time–jerk optimal trajectory planning for industrial robots with coupled interpolation function selection

AbstractIn the contemporary field of optimal trajectory planning for industrial robots, it is customary to construct trajectories through the manual predefinition of interpolation functions. Unfortunately, this method frequently overlooks the influence of the interpolation function itself on the optimization objectives, resulting in suboptimal outcomes. To remedy this limitation, an optimal trajectory planning method with coupled interpolation function selection is proposed, in which the total task time and the integral squared jerk are defined as optimization objectives. This method minimizes the optimization objectives while also factoring in the optimal interpolation function, and avoiding subjective interference. To address the aforementioned biobjective optimization problem better, an Improved MultiObjective Golden Eagle Optimizer is introduced. Population diversity and the ability to escape local optima are enhanced through the incorporation of Chaotic Mapping, Opposition‐Based Learning, Differential Evolution, and adaptive inertia weight strategy into the algorithm. The superiority of the algorithm is validated through a series of simulations on 17 benchmark functions. In the context of the robotic stirring operation within the automated block cast charging process, the proposed method is utilized to derive the time–jerk optimal trajectory. The results demonstrate the effectiveness of the proposed method.

DP-EPSO: Differential privacy protection algorithm based on differential evolution and particle swarm optimization

In deep learning differential privacy protection, adding noise based on gradient has become a mainstream algorithm, but excessive gradient perturbation noise causes accuracy degradation. To solve this problem, a differential privacy protection algorithm based on differential evolution and particle swarm optimization is proposed to realize hyperparameter optimization in differential privacy, reduce the impact of noise on the model, and effectively improve the accuracy. On the one hand, the differential evolution scheme performs selection, crossover and mutation on learning rate η, make it approach the global optimal solution, and improve the computational efficiency of the algorithm. On the other hand, the particle swarm optimization scheme is adopted. Without changing the parameters and gradient structure, the parameters are optimized by using the network propagation attributes, which reduces the influence of noise on the accuracy. Experiments are performed on three datasets: Cifar10, Mnist and FashionMnist. Compared with the existing differential privacy algorithms, under the same privacy budget, the proposed algorithm has better accuracy and higher efficiency.

Using Optimized Long Short-Term Memory For Time-Series Forecasting of Electric Vehicles Battery Charging

The last decade has seen a significant rise in the adoption and development of Electric Vehicles (EVs), driven by environmental concerns, technological advancements, and governmental support. Batteries, central to EVs, have witnessed groundbreaking innovations in terms of energy density, charging speeds, and longevity. Expanding charging infrastructure and the automotive industry's investment in EV research have made them more mainstream. Effective Battery Management (BM), which includes monitoring essential parameters and thermal management, is critical for the longevity and reliability of EVs. Accurate charge prediction, in particular, aids in trip planning, reduces range anxiety and facilitates cost-effective charging coordinated with dynamic electricity pricing. Traditional models like linear regression and Autor-Rgressive Integrated Moving Average (ARIMA) have been standard for EV battery charge prediction. However, these often struggle with the dynamic nature of EV charging data. Even models like the vanilla Long Short-Term Memory (LSTM), which are adept at recognizing long-term patterns, require meticulous hyperparameter tuning. This work introduces the DWT-DE-LSTM model, which utilizes the Discrete Wavelet Transform (DWT) to dissect battery charging data at different resolutions and a Differential Evolution (DE) strategy for model optimization. Tests using the Panasonic 18650PF Li-ion Battery Dataset revealed the superior efficacy of the DWT-DE-LSTM model, emphasizing its suitability for real-world battery charge prediction.

Blockchain event ordering algorithm based on high-dimensional multi-objective differential evolution

Abstract Since the birth of Bitcoin, blockchain technology has been widely used in many fields, such as finance, supply chain management, smart contracts, etc.and has been highly valued for its decentralized and tamper-proof characteristics. With the expansion of application scenarios, blockchain faces inefficient data processing challenges, particularly in event sequencing and processing. Developing efficient event sequencing algorithms is a research priority to enhance the performance of blockchain systems. This study addresses the efficiency problem in blockchain event sequencing by introducing a high-dimensional multi-objective differential evolution strategy to improve blockchain technology’s performance in various fields. This paper uses the NSGA-III-WA algorithm to efficiently sort blockchain events by combining differential evolution strategies. Compared to the traditional sorting algorithm, this algorithm reduces latency by 30% and improves memory usage by about 20% when dealing with events of the same size, ensuring sorting accuracy. By optimizing the weight vector adjustment strategy, the convergence speed and distribution uniformity of the algorithm and the solution can be optimized. In this paper, the algorithm can effectively improve the efficiency and performance of blockchain event processing, which is critical in optimizing blockchain systems.

A multi-strategy improved snake optimizer and its application to SVM parameter selection

<p>Support vector machine (SVM) is an effective classification tool and maturely used in various fields. However, its performance is very sensitive to parameters. As a newly proposed swarm intelligence algorithm, snake optimizer algorithm (SO) can help to solve the parameter selection problem. Nevertheless, SO has the shortcomings of weak population initialization, slow convergence speed in the early stage, and being easy to fall into local optimization. To address these problems, an improved snake optimizer algorithm (ISO) was proposed. The mirror opposition-based learning mechanism (MOBL) improved the population quality to enhance the optimization speed. The novel evolutionary population dynamics model (NEPD) was beneficial for searching accurately. The differential evolution strategy (DES) helped to reduce the probability of falling into local optimal value. The experimental results of classical benchmark functions and CEC2022 showed that ISO had higher optimization precision and faster convergence rate. In addition, it was also applied to the parameter selection of SVM to demonstrate the effectiveness of the proposed ISO.</p>

An improved composite particle swarm optimization algorithm for solving constrained optimization problems and its engineering applications

<abstract><p>In the last few decades, the particle swarm optimization (PSO) algorithm has been demonstrated to be an effective approach for solving real-world optimization problems. To improve the effectiveness of the PSO algorithm in finding the global best solution for constrained optimization problems, we proposed an improved composite particle swarm optimization algorithm (ICPSO). Based on the optimization principles of the PSO algorithm, in the ICPSO algorithm, we constructed an evolutionary update mechanism for the personal best position population. This mechanism incorporated composite concepts, specifically the integration of the $ \varepsilon $-constraint, differential evolution (DE) strategy, and feasibility rule. This approach could effectively balance the objective function and constraints, and could improve the ability of local exploitation and global exploration. Experiments on the CEC2006 and CEC2017 benchmark functions and real-world constraint optimization problems from the CEC2020 dataset showed that the ICPSO algorithm could effectively solve complex constrained optimization problems.</p></abstract>

Optimal wind power generation system by honey badger algorithm with differential evolution strategies

The purpose of this study is to build an optimal hybrid wind power system consisting of a permanent magnet direct-drive wind power generation unit, a hybrid energy storage system (HESS), a power electronic converter, and loads. Moreover, a reasonable control method is designed for each part, and a honey badger algorithm (HBA) with differential evolution strategies is employed to realize the coordinated control of each unit and improve the system stability. The real anti-interference capability of wind power system can be improved by the proposed HBA with differential evolution strategies (IHBA). Firstly, the wind hybrid system model is constructed, in which the wind power generation unit adopts the variable step climbing method to achieve the maximum power point tracking control; the HESS designs the power distribution method based on the bus voltage; the converter supplying energy to the AC load introduces the virtual synchronous generator (VSG) control strategy. Secondly, due to the existence of energy exchange in each unit of the system, VSG has more parameters and its control performance is influenced by the system itself, this study proposes HBA with differential evolution strategies for system parameter optimization and constructs an IHBA-VSG model with the objective of minimizing the bus voltage fluctuation value. The IHBA improves original honey badger algorithm by three mechanisms which are time control function, Gaussian variation factor and differential evolution strategy. Finally, the parameters obtained from the HBA with differential evolution strategies optimization search are substituted into the wind hybrid system, and the simulation system is built by MATLAB/Simulink. The simulation results show that when the environment of the system changes or the load on the AC side changes abruptly, the proposed control method can reduce the bus voltage fluctuation by 2.9%− 18.22% compared with the honey badger algorithm Optimized VSG (HBA-VSG), and can reduce the bus voltage fluctuation by 5.09%− 17.98% compared with the traditional droop control without considering the system interaction. This study can effectively improve the stability of wind power generation system and promote the development of new energy industry.

An equilibrium honey badger algorithm with differential evolution strategy for cluster analysis

Data clustering is a machine learning method for unsupervised learning that is popular in the two areas of data analysis and data mining. The objective is to partition a given dataset into distinct clusters, aiming to maximize the similarity among data objects within the same cluster. In this paper, an improved honey badger algorithm called DELHBA is proposed to solve the clustering problem. In DELHBA, to boost the population’s diversity and the performance of global search, the differential evolution method is incorporated into algorithm’s initial step. Secondly, the equilibrium pooling technique is included to assist the standard honey badger algorithm (HBA) break free of the local optimum. Finally, the updated honey badger population individuals are updated with Levy flight strategy to produce more potential solutions. Ten famous benchmark test datasets are utilized to evaluate the efficiency of the DELHBA algorithm and to contrast it with twelve of the current most used swarm intelligence algorithms and k-means. Additionally, DELHBA algorithm’s performance is assessed using the Wilcoxon rank sum test and Friedman’s test. The experimental results show that DELHBA has better clustering accuracy, convergence speed and stability compared with other algorithms, demonstrating its superiority in solving clustering problems.

A Novel Parameter Estimation Scheme for Vehicle Suspension Systems Based on Response and Test Track Prioritization

In this paper, a system identification methodology based on vehicle response and test track prioritization is presented. The proposed method can be used to perform time-domain parameter estimation by simply driving the vehicle on different road profiles, thus eliminating the necessity of having a special test setup or excitation input feed. The method is based on exploiting the dependence of the identifiability of vehicle parameters on vehicle responses and test tracks. To prioritize the test tracks, the responses and test tracks are weighed based on their capabilities to estimate the vehicle parameters, and these weights are obtained in a preliminary analysis of a seven-degree-of-freedom ride model developed in MATLAB® (R2021b). An optimization framework based on a differential evolution scheme with an optional archive scheme (JADE) is used to estimate the twelve vehicle parameters. The proposed system identification scheme is verified on tests using a high-fidelity vehicle model developed in ADAMS® (29.2)using several tracks with different severities. Experimental validation is performed on a test vehicle, and it is demonstrated that the vehicle suspension system parameters can be identified accurately with fast convergence performance.

Flood susceptibility mapping using AutoML and a deep learning framework with evolutionary algorithms for hyperparameter optimization

Flooding is one of the most common natural hazards that have extremely detrimental consequences. Understanding which areas are vulnerable to flooding is crucial to addressing these effects. In this work, we use machine learning models and Automated machine learning (AutoML) systems for flood susceptibility mapping in Kerala, India. In particular, we used a three-dimensional convolutional neural network (CNN) architecture for this purpose. The CNN model was assisted with hyperparameter optimization techniques that combine Bayesian optimization with evolutionary algorithms like differential evolution and covariance matrix adaptation evolutionary strategies. The performances of all models are compared in terms of cross-entropy loss, accuracy, precision, recall, area under the curve (AUC) and kappa score. The CNN model shows better performance than the AutoML models. Evolutionary algorithm-assisted hyperparameter optimization methods improved the efficiency of the CNN model by 4 and 9 percent in terms of accuracy and by 0.0265 and 0.0497 with reference to the AUC score.