Tuning Algorithm Research Articles

Reinforcement learning-trained optimisers and Bayesian optimisation for online particle accelerator tuning

Online tuning of particle accelerators is a complex optimisation problem that continues to require manual intervention by experienced human operators. Autonomous tuning is a rapidly expanding field of research, where learning-based methods like Bayesian optimisation (BO) hold great promise in improving plant performance and reducing tuning times. At the same time, reinforcement learning (RL) is a capable method of learning intelligent controllers, and recent work shows that RL can also be used to train domain-specialised optimisers in so-called reinforcement learning-trained optimisation (RLO). In parallel efforts, both algorithms have found successful adoption in particle accelerator tuning. Here we present a comparative case study, assessing the performance of both algorithms while providing a nuanced analysis of the merits and the practical challenges involved in deploying them to real-world facilities. Our results will help practitioners choose a suitable learning-based tuning algorithm for their tuning tasks, accelerating the adoption of autonomous tuning algorithms, ultimately improving the availability of particle accelerators and pushing their operational limits.

Simulation and Arduino Hardware Implementation of ACO, PSO, and FPA Optimization Algorithms for Speed Control of a DC Motor

This article proposes implementing and comparing the effectiveness of three optimization algorithms (ACO, PSO, and FPA) for tuning a proportional-integral-derivative (PID) controller on an Arduino Mega 2560 board. This relatively unexplored approach aims to evaluate these algorithms through practical experiments. The choice of PID control is due to its design simplicity and widespread industrial use. Similarly, the permanent magnet DC motor (PMDC) was selected because of its crucial role in various industrial sectors. Tuning PID parameters using optimization algorithms has garnered increasing interest due to its demonstrated efficiency. Several studies have validated the stability of ACO, PSO, and FPA algorithms, justifying their selection. In this article, simulation results showed that ACO, with a response time of 0.322s and an overshoot of 0.68%, was more effective than PSO, which had a response time of 0.768s and an overshoot of 13%. FPA had a response time of 0.347s, close to ACO, but a higher overshoot of 6%. In practice, several factors come into play, such as speed ripples caused by the speed sensor, and machine saturation, which must be considered to ensure practical implementation. After adjusting the PID parameters and integrating a low-pass filter in the feedback loop, ACO, with a response time of 0.596s and an overshoot of 1.68%, was very close to FPA, which had a response time of 0.644s and an overshoot of 0.81%. This comparison highlighted the advantages of the FPA algorithm, which is simple to use, requires fewer parameters to adjust, and takes less time than ACO. This study suggests the potential for implementing a hybrid FPA-ACO algorithm, leveraging the strengths of both algorithms.

Enhancing Network Anomaly Intrusion Detection with IoT Data-Driven BOA-CNN-BiGRU-AAM -Net Classification

Network security is one of the key components of cybersecurity anomaly intrusion detection, which is responsible for identifying unusual behaviours or activities within a network that might indicate possible security breaches or threats. In this suggested intrusion detection system (IDS), network traffic data is continuously monitored via anomaly detection. The study makes utilising one of the most recent datasets to spot unusual behaviour in networks connected to the Internet of Things, the IoTID20 dataset, to facilitate this process. The preprocessing stage involves painstaking steps for smoothing, filtering, and cleaning the data. The Pine Cone Optimisation algorithm (PCOA), a novel optimizer inspired by nature, is introduced in this study for the feature selection process. PCOA seeks to increase the effectiveness of feature selection while drawing inspiration from the various ways that pine trees reproduce, such as pollination and the movement of pine cones by animals and gravity. Moreover, IDS is classified using Bidirectional Gated Recurrent Unit–Additive Attention Mechanism Based on Convolutional Neural Networks (CNN-BiGRU-AAM), which makes use of deep learning's capabilities for efficient classification tasks. In addition, this work presents the Botox Optimisation Algorithm (BOA) for hyperparameter tuning, which is modelled after the way Botox functions in human anatomy. BOA uses a human-based method to adjust the hyperparameters of the model to attain the best accuracy. The results of the experiments show that the suggested methodologies are effective in improving network anomaly intrusion detection systems, with a maximum accuracy of 99.45%.

Current Harmonic Suppression in Maritime Vessel Rudder PMSM Drive System Based on Composite Fractional-Order PID Repetitive Controller

To address the control performance and harmonic suppression issues in maritime vessel rudder permanent magnet servo systems, a fractional-order PID controller was introduced into the existing improved repetitive control strategy. We used the Oustaloup approximation algorithm and particle swarm optimization for tuning the fractional-order PID controller. The optimized parameters substantially improved the control performance. By integrating the fractional-order PID controller with the improved repetitive controller, a composite fractional-order PID repetitive control strategy was formed. Finally, MATLAB/Simulink simulations were conducted to compare and verify the disturbance rejection and harmonic suppression capabilities of the improved control strategy. The results demonstrate its superior control performance, thereby increasing the practicality of the control system in dealing with various situations.

Improved Dujiangyan Irrigation System Optimization (IDISO): A Novel Metaheuristic Algorithm for Hydrochar Characteristics

Hyperparameter tuning is crucial in the development of machine learning models. This study introduces the nonlinear shrinking factor and the Cauchy mutation mechanism to improve the Dujiangyan Irrigation System Optimization (DISO), proposing the improved Dujiangyan Irrigation System Optimization algorithm (IDISO) for hyperparameter tuning in machine learning. The optimization capabilities and convergence performance of IDISO were validated on 87 CEC2017 benchmark functions of varying dimensions and nine real-world engineering problems, demonstrating that it significantly outperforms DISO in terms of convergence speed and accuracy, and ranks first in overall performance among the seventeen advanced metaheuristic algorithms being compared. To construct a robust and generalizable prediction model for hydrochar element characteristics, this study utilized IDISO and DISO algorithms to fine-tune the parameters of the XGBoost model. The experimental results show that the IDISO-XGBoost model achieved an average prediction performance of 0.95, which represents a 4% improvement over the DISO-XGBoost model. These results indicate that the IDISO algorithm has significant potential and value in practical applications.

A practical parameter tuning algorithm for super-twisting algorithm-based differentiator and its application in altitude ground test facility

The super-twisting algorithm (STA)-based robust exact differentiator (RED) proposed by Levant has been widely applied in control, observation, and fault detection. With properly set parameters, STA-RED is able to estimate exactly and robustly the derivatives of an input. With a given input signal, however, a practical parameter tuning rule still lacks. To this end, a general parameter tuning algorithm with the full stability proof for STA-RED is proposed. In particular, the time linear transformation and the coordinate transformation are performed on a given base signal, allowing the deterministic relationship between the parameters of STA-RED and the amplitude and the frequency of the input to be obtained. The full stability of the STA-RED based on the proposed parameter tunning algorithm (denoted as PSTA-RED) with perturbations is given using the Lyapunov function method. The simulations are carried out using given input signals and the pressure signals from the altitude ground test facility to illustrate the effectiveness of the proposed PSTA-RED.

Data‐driven predictive control of perturbed buck converters using a modified iterative feedback tuning algorithm

AbstractThe most challenging aspect of utilizing model predictive controllers (MPCs), particularly those involving power electronic applications, is the extraction of a model that accurately represents the behavior of the studied system. Concerning the use of power electronic applications, as long as an MPC is used, adjusting the controller parameters brings difficulties. In addition, as the number of elements increases, it becomes harder to get the best control law out of the model. To do away with the need for model extraction, this study presents an offline data‐driven approach in conjunction with the MPC that can optimally adjust the MPC parameters based on the iterative feedback tuning (IFT) algorithm called the iterative feedback predictive controller (IFPC). The proposed method eliminates concerns regarding selecting an optimal number of algorithm iterations, thereby reducing operating costs, by introducing a modified IFT called feedback‐based IFPC (FIFPC) while simultaneously achieving optimal MPC parameters. The proposed method is applied to a constant voltage load (CVL) connected less‐than‐ideal buck converter, that is, one with perturbed filter elements and variable loads. A robust stability analysis (RSA) is performed under normal operating conditions to investigate the robustness behavior of the proposed controller. Simulation studies are presented to evaluate the proposed controller under different scenarios, such as step and abrupt load changes and measurement noise, compared with the well‐known model‐based and data‐enabled predictive controller (DeePC) approaches in the MATLAB/Simulink environment.

A New Tuning Algorithm Based on Variable Step Approaching for RF Self-Interference Cancellation

A New Tuning Algorithm Based on Variable Step Approaching for RF Self-Interference Cancellation

Revolutionizing connectivity: Unleashing the power of 5G wireless networks enhanced by artificial intelligence for a smarter future

In the context of 5G wireless networks, this paper addresses a crucial concern—data loss in Analog-to-Digital Converters (ADCs) embedded within Multiple Input Multiple Output (MIMO) channels, a prevalent challenge in modern wireless communication systems. Despite the integration of ADCs in MIMO configurations to mitigate this issue, their implementation introduces significant complexities. In response to this, the paper introduces an innovative and efficient approach aimed at enhancing channel estimation in 5G MIMO systems. The proposed methodology relies on a heuristic-based optimization technique within a channel prediction framework. The framework uses feedback information collected through a Cascade Fusion Serial Network (CFSN) to analyze the receiver-side error ratio and estimate the channel coefficients of MIMO systems at the transmitter. An attention mechanism is incorporated into the model design through the use of a Convolutional Neural Networks - Gated Recurrent Units (CNN-GRU) and a Multi-Scaled Stacked Auto encoder. Parameters of Attention CNN-GRU, Stacked Auto encoder, and CFSN are refined using the Whale Optimization method. The optimization objective is to minimize key metrics, including Root Mean Squared Error (RMSE), Bit Error Rate (BER), and Mean Squared Error (MSE) associated with the estimated channel. Experimental evaluations underscore the efficacy of the proposed MIMO model in the context of 5G networks. The results demonstrate improved convergence rates, heightened prediction performance, and reduced computational costs compared to existing methods. This research contributes valuable insights into addressing data loss challenges specific to 5G MIMO systems, providing a pioneering solution that integrates heuristic optimization, advanced neural network architectures, and nature-inspired algorithms for parameter tuning.

Forecasting U.S. Fossil Energy Consumption: Advancing Accuracy with Multilayer Perceptron and Residual Learning

Accurate midterm forecasts of fossil fuel consumption are essential for effective energy planning, economic management, and resource allocation. While machine learning models have demonstrated their efficacy in handling large-scale nonlinear datasets, many, including Multilayer Perceptrons (MLPs), suffer from performance degradation with increased depth. Fortunately, recent studies have revealed that Residual Networks (ResNets) can mitigate or even overcome this challenge. In this paper, we propose a Weighted Residual Network based on MLP to enhance predictive performance. We employ the Adam algorithm for model training and utilize the Gridsearch algorithm for hyperparameter tuning. In the application section, we develop predictive models using three case studies: natural gas, petroleum, and total fossil fuel consumption. We validate the effectiveness of our proposed model and compare it with ten other machine learning models. Our findings demonstrate that our proposed model consistently outperforms others in all three cases, underscoring its superior performance in midterm forecasting of fossil fuel consumption.

Addressing class-imbalanced learning in real-time aero-engine gas-path fault diagnosis via feature filtering and mapping

Condition-based maintenance of aero-engines requires real-time gas-path fault diagnosis, which is crucial for reducing costs and enhancing aircraft attendance. It is imperative to ensure a low false alarm rate in fault diagnosis algorithms. Inherent variations between engines and age-related performance deterioration can lead to an increase in false alarms, necessitating online algorithm tuning. However, insufficient fault data in online tuning may lead to class imbalance and algorithmic forgetfulness of fault features, compromising effectiveness. To tackle this issue, this paper introduces a novel hybrid diagnosis algorithm that utilizes feature filtering and mapping. This method entails an engine model, a feature extractor, a feature filter, a fault classifier, and several feature mappers. The feature mappers are designed to generate fault features based on the incoming fault free data for rebalancing the tuning dataset. And the feature filter is introduced to prevent weight collapse during the online tuning. Validation through a 400-hour life cycle simulation test and a real data test indicate that the method outperforms the benchmark methods. It overcomes the class-imbalanced learning problem, showcasing its potential for aero-engine gas-path fault diagnosis.

Objective Evaluation of Motion Cueing Algorithms for Vehicle Driving Simulator Based on Criteria Importance through Intercriteria Correlation (CRITIC) Weight Method Combined with Gray Correlation Analysis

Perception-based fidelity evaluation metrics are crucial in driving simulators, as they play a key role in the automatic tuning, assessment, and comparison of motion cueing algorithms. Nevertheless, there is presently no unified and effective evaluation framework for these algorithms. To tackle this challenge, our study initially establishes a model rooted in visual–vestibular interaction and head tilt angle perception systems. We then employ metrics like the Normalized Average Absolute Difference (NAAD), Normalized Pearson Correlation (NPC), and Estimated Delay (ED) to devise an evaluation index system. Furthermore, we use a combined approach incorporating CRITIC and gray relational analysis to ascertain the weights of these indicators. This allows us to consolidate them into a comprehensive evaluation metric that reflects the overall fidelity of motion cueing algorithms. Subjective evaluation experiments validate the reasonableness and efficacy of our proposed Perception Fidelity Evaluation (PFE) method.

Optimizing power system stability: A hybrid approach using manta ray foraging and Salp swarm optimization algorithms for electromechanical oscillation mitigation in multi‐machine systems

AbstractThis paper emphasizes the significance of ensuring adequate damping of electromechanical oscillations in power systems to ensure stable operation. Power System Stabilizers (PSSs) are influential in enhancing system damping and refining dynamic characteristics during transient conditions. However, the efficacy of PSSs is notably contingent on parameter values, particularly in the case of lead‐lag PSSs. In response to this challenge, the paper introduces a Tilt‐Integral‐Derivative (TID)‐based PSS, optimized through a novel optimization algorithm called Hybrid Manta Ray Foraging and Salp Swarm Optimization Algorithms (MRFOSSA). The MRFOSSA algorithm demonstrates robustness and enhanced convergence, validated through benchmark function tests, and outperforms competing algorithms. These superior characteristics of MRFOSSA were employed in optimal tuning of TID‐PSSs to uphold the stability of multi‐machine power systems. The MRFOSSA algorithm demonstrates robustness and enhanced convergence, outperforms competing algorithms in the optimal tuning of TID‐PSS within the Western System Coordinating Council (WSCC)‐3‐machines 9‐bus test system. In summary, the proposed TID‐PSS, coupled with the MRFOSSA algorithm, presents a promising avenue for enhancing power system stability.

Tuning-free sparse clustering via alternating hard-thresholding

Model-based clustering is a commonly-used technique to partition heterogeneous data into homogeneous groups. When the analysis is to be conducted with a large number of features, analysts face simultaneous challenges in model interpretability, clustering accuracy, and computational efficiency. Several Bayesian and penalization methods have been proposed to select important features for model-based clustering. However, the performance of those methods relies on a careful algorithmic tuning, which can be time-consuming for high-dimensional cases. In this paper, we propose a new sparse clustering method based on alternating hard-thresholding. The new method is conceptually simple and tuning-free. With a user-specified sparsity level, it efficiently detects a set of key features by eliminating a large number of features that are less useful for clustering. Based on the selected key features, one can readily obtain an effective clustering of the original high-dimensional data under a general sparse covariance structure. Under mild conditions, we show that the new method leads to clusters with a misclassification rate consistent to the optimal rate as if the underlying true model were used. The promising performance of the new method is supported by both simulated and real data examples.

LSTM-Autoencoder Deep Learning Model for Anomaly Detection in Electric Motor

Anomaly detection is the process of detecting unusual or unforeseen patterns or events in data. Many factors, such as malfunctioning hardware, malevolent activities, or modifications to the data’s underlying distribution, might cause anomalies. One of the key factors in anomaly detection is balancing the trade-off between sensitivity and specificity. Balancing these trade-offs requires careful tuning of the anomaly detection algorithm and consideration of the specific domain and application. Deep learning techniques’ applications, such as LSTMs (long short-term memory algorithms), which are autoencoders for detecting an anomaly, have garnered increasing attention in recent years. The main goal of this work was to develop an anomaly detection solution for an electrical machine using an LSTM-autoencoder deep learning model. The work focused on detecting anomalies in an electrical motor’s variation vibrations in three axes: axial (X), radial (Y), and tangential (Z), which are indicative of potential faults or failures. The presented model is a combination of the two architectures; LSTM layers were added to the autoencoder in order to leverage the LSTM capacity for handling large amounts of temporal data. To prove the LSTM efficiency, we will create a regular autoencoder model using the Python programming language and the TensorFlow machine learning framework, and compare its performance with our main LSTM-based autoencoder model. The two models will be trained on the same database, and evaluated on three primary points: training time, loss function, and MSE anomalies. Based on the obtained results, it is clear that the LSTM-autoencoder shows significantly smaller loss values and MSE anomalies compared to the regular autoencoder. On the other hand, the regular autoencoder performs better than the LSTM, comparing the training time. It appears then, that the LSTM-autoencoder presents a superior performance although it was slower than the standard autoencoder due to the complexity of the added LSTM layers.

Design and Implementation of PID Controller for the Cooling Tower's pH Regulation Based on Particle Swarm Optimization PSO Algorithm

The PH regulation of cooling tower plant in southern fertilizers company (SCF) in Iraq is important for industry pipes protection and process continuity. According to the Mitsubishi standard, the PH of cooling water must be around (7.1 to 7.8). The deviation in PH parameter affects the pipes, such as corrosion and scale. Acidic water causes pipes to corrode, and alkaline water causes pipes to scale. The sulfuric acid solution is used for PH neutralization. The problem is that the sulfuric acid is pumped manually in the cooling tower plant every two or three hours for PH regulation. The manual operation of the sulfuric acid pump makes deviations in the PH parameter. It is very difficult to control the PH manually. To solve this problem, a PID controller for PH regulation was used. The reason for using the PID controller is that the PH response is irregular through the neutralization process. The methodology is to calculate the transfer function of the PH loop using the system identification toolbox of MATLAB, to design and implement a PID controller, to optimize the PID controller response using particle swarm optimization PSO algorithm, and to make a comparison among several tuning methods such as Ziegler Nichols (ZN) tuning method, MATLAB tuner method, and PSO algorithm tuning method. The results showed that the PSO-based PID controller tuning gives a better overshoot, less rise time, and an endurable settling time than the other tuning methods. Hence, the PH response became according to the target range. The experimental results showed that the PH regulation improved using the PSO-based PID controller tuning.

Solution path algorithm for distributionally robust regression

ABSTRACT In this paper, we propose a general distributionally robust regression model based on distributionally robust optimization theory. The proposed model has a piecewise linear loss function and elastic net penalty term, and it generalizes many other regression models. We prove the piecewise linear property of the optimal solutions to this model, which enables us to develop a solution path algorithm for the hyperparameter tuning. A Doubly regularized Least Absolute Deviations (DrLAD) regression model is proposed based on this framework, and a solution path algorithm is developed to speed up the tuning of two hyperparameters in this model. Numerical experiments are implemented to validate the performance of this model and the computational efficiency of the solution path algorithm.

An energy-efficient GMRES–multigrid solver for space-time finite element computation of dynamic poroelasticity

We present and analyze computationally Geometric MultiGrid (GMG) preconditioning techniques for Generalized Minimal RESidual (GMRES) iterations to space-time finite element methods (STFEMs) for a coupled hyperbolic–parabolic system modeling, for instance, flow in deformable porous media. By using a discontinuous temporal test basis, a time marching scheme is obtained. Higher order approximations that offer the potential to inherit most of the rich structure of solutions to the continuous problem on computationally feasible grids increase the block partitioning dimension of the algebraic systems, comprised of generalized saddle point blocks. Our V-cycle GMG preconditioner uses a local Vanka-type smoother. Its action is defined in an exact mathematical way. Due to nonlocal coupling mechanisms of 348 unknowns, the smoother is applied on patches of elements. This ensures damping of higher order error frequencies. By numerical experiments of increasing complexity, the efficiency of the solver for STFEMs of different polynomial order is illustrated and confirmed. Its parallel scalability is analyzed. Beyond this study of classical performance engineering, the solver’s energy efficiency is investigated as an additional and emerging dimension in the design and tuning of algorithms on the hardware.

Investigation of Legendre polynomials expanded adaptive controller with inverse compensation for high frequency tracking of electro-hydraulic force systems

Accurate force tracking of an electro-hydraulic force system (EHFS) is of great significance in various industrial applications. As the EHFS is inevitably confronted with hydraulic nonlinearities, dynamic variations and uncertain disturbances, the real-time force tracking performance is generally unsatisfactory, especially for high frequency command force signals. For reducing the force tracking error with relatively high frequency inputs, a Legendre polynomials expanded adaptive controller with inverse compensation is proposed in this paper. The proposed controller is constructed by introducing an offline designed inverse compensation (IC) controller and an online adaptive controller to the EHFS governed by traditional Proportional Integral (PI) controller, in which the former IC controller is acting as the inner loop controller and the latter adaptive controller is regarded as an outer controller. The inner loop IC controller with fixed control parameters is offline designed on the basis of traditional PI controller by means of the generalized projection identification algorithm and the zero magnitude error tracking technology, focusing on extending the real-time force tracking bandwidth of the EHFS. The outer loop online adaptive controller is developed by Legendre polynomials expanded adaptive filters with nonlinear mapping ability, whose weights are online updated by an adaptive tuning algorithm with the aim of handling system’s dynamic variations, nonlinearities and uncertain disturbances. The proposed controller is implemented on a real EHFS test rig by the xPC/Target rapid prototyping technique, and comparative experimental results demonstrate that the proposed controller can achieve much higher high frequency tracking performance than the traditional PI and IC controller commonly used in industry.

Active Damping for Dynamic Improvement of Multiple Grid-Tied Virtual Synchronous Generators

To eliminate low-frequency oscillations, this paper proposes an active-damping method for multiple grid-tied virtual synchronous generators (VSGs) in a power plant. Firstly, using the Lyapunov's indirect method, the damping ratio of multiple VSGs in parallel is analyzed. The average damping ratio reveals that this multi-VSG power plant can be poorly damped in a wide range of inertia and damping settings. Then, self- and mutual-damping controllers are developed to suppress self- and mutually induced low-frequency power oscillations, respectively. For the practical implementation, an adaptive tuning algorithm that enables automatic realization is proposed. Through a reassessment, a remarkable damping-ratio improvement is validated. Moreover, the inertial response improvement is validated by the frequency response analysis. Finally, simulations in Digsilent/PowerFactory and experiments are performed to demonstrate the accuracy of the analyses and the effectiveness of the proposed method.