Adaptive Algorithm Research Articles

Reducing measurement costs by recycling the Hessian in adaptive variational quantum algorithms

Abstract Adaptive protocols enable the construction of more efficient state preparation circuits in variational quantum algorithms (VQAs) by utilizing data obtained from the quantum processor during the execution of the algorithm. This idea originated with Adaptive Derivative-Assembled Problem-Tailored variational quantum eigensolver (ADAPT-VQE), an algorithm that iteratively grows the state preparation circuit operator by operator, with each new operator accompanied by a new variational parameter, and where all parameters acquired thus far are optimized in each iteration. In ADAPT-VQE and other adaptive VQAs that followed it, it has been shown that initializing parameters to their optimal values from the previous iteration speeds up convergence and avoids shallow local traps in the parameter landscape. However, no other data from the optimization performed at one iteration is carried over to the next. In this work, we propose an improved quasi-Newton optimization protocol specifically tailored to adaptive VQAs. The distinctive feature in our proposal is that approximate second derivatives of the cost function are recycled across iterations in addition to optimal parameter values. We implement a quasi-Newton optimizer where an approximation to the inverse Hessian matrix is continuously built and grown across the iterations of an adaptive VQA. The resulting algorithm has the flavor of a continuous optimization where the dimension of the search space is augmented when the gradient norm falls below a given threshold. We show that this inter-optimization exchange of second-order information leads the approximate Hessian in the state of the optimizer to be consistently closer to the exact Hessian. As a result, our method achieves a superlinear convergence rate even in situations where the typical implementation of a quasi-Newton optimizer converges only linearly. Our protocol decreases the measurement costs in implementing adaptive VQAs on quantum hardware as well as the runtime of their classical simulation.

Optimum solution of power flow problem based on search and rescue algorithm

AbstractIn order to solve the optimal power flow (OPF) problem, a unique algorithm based on a search and rescue method is applied in this study. For the OPF problem under three objective functions, the SAR offers a straightforward and reliable solution. The three objective functions are used to minimize the fuel cost, power loss and voltage deviation as a single objective function. The OPF problem for benchmark test system, including the IEEE-14 bus, IEEE-30 bus, and IEEE-57 bus, are solved by the Search and Rescue algorithm (SAR) under specific objective functions that are determined by the operational and economic performance indices of the power system. To demonstrate the efficacy and possibilities of the SAR algorithm, SAR is contrasted with alternative optimization techniques such as harmony search algorithm, gradient method, adaptive genetic algorithm, biogeography-based optimization, Artificial bee colony, gravitational search algorithm, particle swarm optimization, Jaya algorithm, enhanced genetic algorithm, modified shuffle frog leaping algorithm, practical swarm optimizer, Moth flam optimizer, whale and moth flam optimizer, grey wolf optimizer, cheap optimization algorithm and differential evolution algorithm. The value of minimum power losses based on SAR technique is equal to 0.459733441487247 MW for IEEE-14 bus. The value of minimum total fuel cost based on SAR technique is equal to 8051.12225602148 $/h for IEEE-14 bus. The value of minimum voltage deviation based on SAR technique is equal to 0.0357680148269292 for IEEE-14 bus. The value of minimum power losses based on SAR technique is equal to 2.71286428848434 MW for IEEE-30 bus. The value of minimum total fuel cost based on SAR technique is equal to 798.197578585806 $/h for IEEE-30 bus. The value of minimum voltage deviation based on SAR technique is equal to 0.0978069572088536 for IEEE-30 bus. The value of minimum total fuel cost based on SAR technique is equal to 38017.7691758245 $/h for IEEE-57 bus. The acquired results for the OPF compared to all competitor algorithms in every case of fitness function demonstrate the superiority of the SAR method.

PID control algorithm based on multistrategy enhanced dung beetle optimizer and back propagation neural network for DC motor control.

Traditional Proportional-Integral-Derivative (PID) control systems often encounter challenges related to nonlinearity and time-variability. Original dung beetle optimizer (DBO) offers fast convergence and strong local exploitation capabilities. However, they are limited by poor exploration capabilities, imbalance between exploration and exploitation phases, and insufficient precision in global search. This paper proposes a novel adaptive PID control algorithm based on enhanced dung beetle optimizer (EDBO) and back propagation neural network (BPNN). Firstly, the diversity of exploration is increased by incorporating a merit-oriented mechanism into the rolling behavior. Then, a sine learning factor is introduced to balance the global exploration and local exploitation capabilities. Additionally, a dynamic spiral search strategy and adaptive [Formula: see text]-distribution disturbance are presented to enhance search precision and global search capability. The BPNN is employed to fine-tune both PID and network parameters, leveraging its powerful generalization and learning ability to model nonlinear system dynamics. In the simplified motor experiments, the proposed controller achieved the lowest overshoot (0.5%) and the shortest response time (0.012s), with a settling time of 0.02s and a steady-state error of just 0.0010. In another set of experiments, the proposed controller recorded an overshoot and response time of 0.7% and 0.0010s, across five DC motor tests. These results demonstrate the proposed adaptive PID control algorithm has superior performance in optimizing control system parameters, as well as improving system robustness and stability.

Adaptive algorithms for drone flight control under communication constraints and information incompleteness

Abstract With the rapid increase in the use of drones in various applications, including commercial and governmental, and the increasing probability of communication failures and contingencies, research becomes critical to ensure the safety and efficiency of their operations. The aim of this research is to develop adaptive drone flight control algorithms capable of operating effectively under conditions of limited communication and incomplete information to ensure reliable and safe autonomous operation of these systems. The applied methods include computer modelling and simulation, analytical, statistical, functional, deductive and descriptive methods. The study found that the use of performance evaluation methods for complex systems enables the identification of safety and performance criteria for drones, and drone flight control provides basic principles and methods that can be adapted for drones, including autopiloting and navigation. In addition, analyses of satellite communication and navigation prove the need to consider the limitations of this technology when developing drone control algorithms. The combination of these techniques allows for more robust and adaptive drone control systems that can function effectively in complex environments such as communication limitations and incomplete information. Additionally, it was found that the integration of adaptive control algorithms based on these methods allows drones to effectively adapt to variable environmental conditions and make decisions quickly even when communication is lost or information is limited.

Generalized Entropy and Solution Information for Measuring Puzzle Difficulty

Metrics for problem difficulty are used by many puzzle generation algorithms, as well as by adaptive algorithms that want to provide players with the puzzles at the correct level of difficulty. A recently proposed general metric, puzzle entropy, combines an analysis of game mechanics with a model of player knowledge in the form of inference rules to predict problem difficulty. The entropy of a puzzle is the amount of information required, given a player’s knowledge about the puzzle, to describe a solution to a puzzle. This paper generalizes the concepts of puzzle entropy and solution information, providing a better foundation for the previous work and creating new algorithms, Minimum Solution Information and Total Solution Information. While functionally very similar to past work, the new algorithm allows knowledge about a puzzle to be represented as a policy, something that can be learned more easily. We then evaluate the impact of policies, inference rules, and player knowledge in the 2016 game, The Witness.

Study on intelligent recognition of urban road subgrade defect based on deep learning.

China's operational highway subgrades exhibit a trend of diversifying types and an increasing number of defects, leading to more frequent urban road safety incidents. This paper starts from the non-destructive testing of urban road subgrade defects using geological radar, aiming to achieve intelligent identification of subgrade pathologies with geological radar. The GprMax forward simulation software is used to establish multi-layer composite structural models of the subgrade, studying the characteristics of geological radar images for different types of subgrade diseases. Based on the forward simulation images of geological radar for subgrade defects and field measurement data, a geological radar subgrade defect image database is established. The Faster R-CNN deep learning algorithm is applied to achieve target detection, recognition, and classification of subgrade defect images. By comparing the loss value, total number of identified regions, and recognition accuracy as metrics, the study compares four improved versions of the Faster R-CNN algorithm. The results indicate that the faster_rcnn_inception_v2 version is more suitable for the intelligent identification of non-destructive testing of urban road subgrade defects.

Monocular Visual–Inertial Odometer Based on Adaptive Variational Bayes Algorithm

Monocular Visual–Inertial Odometer Based on Adaptive Variational Bayes Algorithm

A Helium Speech Correction Method Based on Generative Adversarial Networks

The distortion of helium speech caused by helium−oxygen gas mixtures significantly impacts the safety and communication efficiency of saturation divers. Although existing correction methods have shown some effectiveness in improving the intelligibility of helium speech, challenges remain in enhancing clarity and high−pitch correction. To address the issue of degraded speech quality post−correction, a novel helium speech correction method based on generative adversarial networks (GANs) is proposed. Firstly, a new helium speech dataset is introduced, which includes isolated words and continuous speech in both Chinese and English. By training and testing on both isolated words and continuous passages, the correction capability of the model can be accurately evaluated. Secondly, a new evaluation system for helium speech correction is proposed, which partially fills the gap in current helium speech evaluation metrics. This system uses comprehensive similarity to evaluate the similarity of keywords at the sentence level, thus assessing the correction results of helium speech from both word and sentence dimensions. Lastly, a GAN−based helium speech correction method is designed. This method solves the problems of pitch period distortion and formant shift in helium speech by introducing an adaptive speech segmentation algorithm and a fusion loss function and significantly improves the clarity and intelligibility of corrected helium speech. The experimental results show that the corrected helium speech is improved in clarity and intelligibility, which shows its practical value and application potential.

Predicting the body weight of crossbred Holstein × Zebu dairy cows using multivariate adaptive regression splines algorithm.

This study aimed to estimate live body weight from body measurements for Holstein × Zebu dairy cows (n = 156) reared under conditions of humid tropics in Mexico using multivariate adaptive regression splines algorithm (MARS) with several train-test proportions. The body measurements included withers height, rump height, hip width, heart girth, body length and diagonal body length. The data were divided into 65:35, 70:30 and 80:20 split data for training and testing sets, respectively. The MARS algorithm was used to construct a prediction model, which predicted the body weight from the body measurements of the test dataset. The results emphasized that the MARS algorithm had an explanation rate for 80:20 train and test set of 0.836 and 0.711, respectively, with minimum Akaike information criterion values. This indicates that it is a reliable way of predicting body weight from body measurements. The results suggest that body weight prediction can be performed with the MARS algorithm in a reliable way, therefore, this algorithm may be a useful tool for animal breeders and researchers in the development of feeding and selection-aimed approaches.

Improved adaptive genetic algorithm for dynamic multi-specification one-dimensional cutting problem

Rebar is an essential material in the construction of bridges and houses. Rebar cutting is an important link in rebar processing, but it is usually completed by manual experience, which is not only time-consuming, but also causes serious waste and reduces the economic benefits. As the country vigorously promotes the green construction method, the traditional rebar cutting method is difficult to meet the development requirements. So dynamic multi-specification one-dimensional cutting problem is studied in this paper. A mathematical model aiming at the maximum utilization rate of raw material is established, and an improved adaptive genetic algorithm is proposed. Large-scale, small-scale, single-specification masterbatch and multi-specification masterbatch are selected for simulation experiments, respectively. The results show that the proposed algorithm can deal with both large- and small-scale multi-specification or single-specification masterbatch cutting problems. Moreover, the algorithm has good performance in solving accuracy and convergence speed, which verifies its feasibility, effectiveness, and stability. Finally, aiming at the problem of dynamic insertion of orders, one-dimensional cutting software is developed, rapid and real-time cutting of rebars is realized, and the utilization rate of building rebars is improved, which plays a positive role in promoting the high-quality development of construction industries such as bridges.

LI-RADS Nonradiation Treatment Response Algorithm Version 2024: Diagnostic Performance and Impact of Ancillary Features.

Background: LI-RADS Treatment Response Algorithm (TRA) version 2024 (v2024) introduced separate algorithms for detecting hepatocellular carcinoma (HCC) viability after radiation and nonradiation locoregional therapies (LRT). The nonradiation algorithm incorporated MRI-based ancillary features to optionally upgrade lesions from LR-TR Equivocal to LR-TR Viable. Objective: To compare the diagnostic performance of LI-RADS Nonradiation TRA v2024 with LI-RADS TRA version 2017 (v2017) and modified RECIST (mRECIST) for evaluating HCC response to LRT on MRI, with attention to the impact of ancillary features. Methods: This retrospective study included 231 patients (198 male, 33 female; median age, 56 years) who underwent LRT for HCC followed by liver resection or transplant between January 2017 and December 2022. Two radiologists (R1, R2) independently evaluated treated lesions (n=306) using LI-RADS Nonradiation TRA v2024, LI-RADS TRA v2017, and mRECIST. Lesions were classified as showing pathologic viability (n=249) or complete pathologic necrosis (n=57) based on curative surgery pathology. Diagnostic performance for pathologic viability was compared using Bonferroni-adjusted McNemar tests, classifying LR-TR Equivocal assessments as test negative. Results: Sensitivity, specificity, and accuracy for LI-RADS Nonradiation TRA v2024 with ancillary features were 85.5%, 75.4%, and 83.7% (R1), and 87.2%, 63.2%, and 82.7% (R2); for LI-RADS Nonradiation TRA v2024 without ancillary features were 81.1%, 78.9%, and 80.7% (R1), and 80.3%, 78.9%, and 80.1% (R2); for LI-RADS TRA v2017 were 79.9%, 82.5%, and 80.4% (R1), and 79.1%, 79.0%, and 79.1% (R2); and for mRECIST, were 83.9%, 54.4%, and 78.4% (R1), and 78.2%, 40.4%, and 78.4% (R2). LI-RADS Nonradiation TRA v2024 with ancillary features showed higher sensitivity and accuracy than LI-RADS Nonradiation v2024 without ancillary features (both readers); higher sensitivity than LI-RADS TRA v2017 (both readers); higher specificity than mRECIST (both readers); and higher accuracy than LI-RADS TRA v2017 (R2) (p<.008); remaining comparisons between LI-RADS Nonradiation TRA v2024 with ancillary features and other systems were not significant (p>.008). Conclusion: LI-RADS Nonradiation TRA v2024 showed good diagnostic performance in detecting pathologic viability. Ancillary features yielded improved sensitivity and accuracy without significant change in specificity. Clinical Impact: Use of LI-RADS Nonradiation TRA v2024 with ancillary features is recommended for guiding prognostic assessments and treatment decisions after LRT.

Automatic face detection based on bidirectional recurrent neural network optimized by improved Ebola optimization search algorithm.

Face detection is a multidisciplinary research subject that employs fundamental computer algorithms, image processing, and patterning. Neural networks, on the other hand, have been widely developed to solve challenges in the domains of feature extraction, pattern detection, and the like in general. The presented study investigates the DNN (deep neural networks) use in the creation of facial detection operating systems. In this study, a novel optimized deep network has been presented to face detection. In this paper, after using some preprocessing stages for contrast enhancement and increasing the data number for the next deep tool, they fed to a bidirectional recurrent neural network (BRNN). The network is optimized via a novel enhanced version of Ebola optimization algorithm to provide far greater accuracy. The suggested procedure is examined on GTFD (Georgia Tech Face Database) and the results indicate that the proposed technique significantly outperforms other comparative methods, attaining an accuracy of 94.3%, a precision of 93.51%, a recall of 94.53%, and an F1-score of 92.47%. Furthermore, the method exhibits resilience against various challenges, achieving an accuracy of 95.6% under occlusions, 96.3% under lighting variations, 94.8% under pose variations, and 92.4% under low resolution conditions. Simulation results depict that the suggested technique gives far greater accuracy in comparison with the other comparative approaches.

Task offloading for multi-server edge computing in industrial Internet with joint load balance and fuzzy security.

The industrial Internet revolutionizes traditional manufacturing through the incorporation of technologies such as real-time production optimization, big data analysis, etc. Computing resource-constrained industrial terminals struggle to effectively execute latency-sensitive and computation-intensive tasks triggered by these technologies. Edge computing (EC) emerges as a promising paradigm for offloading tasks from terminals to the adjacent edge servers, offering the potentiality to augment the computational capacities for industrial terminals. However, the development of accurate offloading strategies poses a prominent challenge for EC in the industrial Internet. Incorrect offloading strategies will misguide the task offloading procedure, resulting in adverse consequences. In this paper, we study the latency-aware multi-server partial EC task offloading problem in the industrial Internet with the consideration of joining load balancing and security protection to provide accurate strategies. Firstly, we establish a task offloading model that supports partial offloading, facilitating latency reduction, task offloading across multiple edge servers with load balance, and accommodation of fuzzy task risks. We quantify the established model as a constrained optimization formulation and prove its NP-hardness. Secondly, to solve the composite offloading strategy comprising both the offloading location and offloading ratio derived from our model, we propose a bi-layer offloading algorithm with joint load balance and fuzzy security, which is based on the adaptive genetic algorithm and simulated annealing particle swarm optimization. Based on extensive experimental results, we find that the established model is effective in reducing the objective value, with a respective decrease of 27% and 46% compared to full execution in edge servers and local execution in industrial terminals. Furthermore, the proposed offloading algorithm exhibits superior performance in terms of solution accuracy compared to existing algorithms.

Advanced hybrid control for induction motor combining ANN-PID speed controller with neural predictive current regulator based on particle swarm optimization

This paper presents a hybrid control strategy for an induction motor (IM) organized into two nested loops. In the outer loop, an adaptive artificial neural network-based proportional-integral-derivative (ANN-PID) controller is used for speed control. The ANN-PID is divided into two stages: the first stage includes an adaptive ANN speed estimator, and the second stage includes an adaptation algorithm that fine-tunes both the weights and biases of the ANN estimator and the parameters of the PID controller. In the internal loop, a predictive current controller is used to control IM currents using neural networks. Specifically, the neural predictive controller (NPC) approaches the control problem by presenting it as an optimization problem using a neural predictor for IM currents. The considered objective function includes two elements: the current tracking error and the electromagnetic torque ripple. This function is computed over a given time horizon informed by predictions of IM currents. The particle swarm optimization (PSO) algorithm is applied to determine the optimal solution for this optimization task. The effectiveness of the hybrid control scheme is demonstrated by simulation results obtained using Matlab/Simulink, highlighting its superiority over conventional control methods. The proposed hybrid control is characterized by a reduction in torque ripple and overshoot, while also showing robustness to variations in load, rotor resistance, and rotor inductance. Comparisons with an ANN controller and a fuzzy PI controller further demonstrate the superior performance of the hybrid controller in handling nonlinear dynamics and maintaining control accuracy under various operating scenarios.

Steered Response Power for Sound Source Localization: a tutorial review.

In the last three decades, the Steered Response Power (SRP) method has been widely used for the task of Sound Source Localization (SSL), due to its satisfactory localization performance on moderately reverberant and noisy scenarios. Many works have analysed and extended the original SRP method to reduce its computational cost, to allow it to locate multiple sources, or to improve its performance in adverse environments. In this work, we review over 200 papers on the SRP method and its variants, with emphasis on the SRP-PHAT method. We also present eXtensible-SRP, or X-SRP, a generalized and modularized version of the SRP algorithm which allows the reviewed extensions to be implemented. We provide a Python implementation of the algorithm which includes selected extensions from the literature.

Numerical Simulation Study on Ice–Water–Ship Interaction Based on FEM-SPH Adaptive Coupling Algorithm

To address the impact of layered ice and seawater on polar vessels navigating in icy waters, this study employs a coupled finite element method (FEM) and smoothed particle hydrodynamics (SPH) algorithm to simulate the collision dynamics between the bow and stern of a designated icebreaker and the ice layers. The foundational principles and deployment strategies of the coupling algorithm have been meticulously delineated, with a subsequent simulation conducted to model the trajectory of icebreakers navigating through stratified ice conditions. The ice load on the hull, the movement of broken ice bodies, and the temporal variation of ice resistance during collision were analyzed. The method’s applicability and precision were substantiated through a comparative analysis between the simulated ice resistance outcomes and the ice load estimations derived from the Lindqvist formula. Finally, the differences between the bow and stern icebreaking methods were compared. The research findings indicate that the coupling algorithm demonstrates high precision in simulating the navigation of icebreakers under layered ice conditions, aligning with actual scenarios. This provides a solid foundation for further exploration of the ice load on polar vessels. Furthermore, at equivalent speeds and ice thicknesses, stern icebreaking was observed to induce greater oscillations in ice load and yield a higher mean resistance compared to bow icebreaking.

Adaptive Prediction Structure for Learned Video Compression

Learned video compression has developed rapidly and shown competitive rate-distortion performance compared with the latest traditional video coding standard H.266 (VVC). However, existing works were restricted to fixed prediction direction and GoP size. The inflexibility on prediction structure hinders learned video compression towards optimal compression efficiency in diverse motion scenarios. In this paper, we propose to advance learned video compression with adaptive prediction structure decision. Specifically, we propose a unified compression framework that supports both forward prediction and bi-directional prediction. The framework can flexibly switch to different prediction direction to achieve better prediction performance. Meanwhile, we propose a low-complexity prediction structure decision algorithm, where prediction direction and GoP size are adaptively determined based on motion complexity to achieve optimal compression efficiency. Experimental results demonstrate that the proposed unified framework with adaptive decision algorithm improves compression efficiency of pure forward prediction-based or bi-directional prediction-based framework with neglectable ( \(0.9\%\) ) encoding time increment. Meanwhile, it achieves comparable compression performance with VVC and recent learned video coding methods.

Adaptive occlusion object detection algorithm based on OL-IoU.

The continuous advancement of autonomous driving technology imposes higher demands on the accuracy of target detection in complex environments, particularly when traffic targets are occluded. Existing algorithms still face significant challenges in detection accuracy and real-time performance under such conditions. To address this issue, this paper proposes an improved YOLOX algorithm based on adaptive deformable convolution, named OCC-YOLOX. This algorithm enhances the feature extraction network's ability to focus on occluded targets by incorporating a coordinate attention mechanism. Additionally, it introduces the Overlapping IoU (OL-IoU) loss function to optimize the overlap between predicted and ground truth bounding boxes, thereby improving detection accuracy. Furthermore, the adoption of Fast Spatial Pyramid Pooling (Fast SPP) reduces computational complexity while maintaining real-time performance. Experiments on fused public datasets demonstrate that OCC-YOLOX achieves improvements in accuracy, recall, and average precision by 2.76%, 1.25%, and 1.92%, respectively. In addition to testing on the KITTI, CityPersons, and BDD100K datasets, the effectiveness of the OCC-YOLOX algorithm is further validated through comparisons with self-collected occlusion scene data. The experimental results indicate that OCC-YOLOX outperforms existing mainstream detection algorithms, particularly in handling complex occlusion scenarios, significantly enhancing the accuracy and efficiency of object detection. This study provides new insights for addressing the challenges of occluded target detection in intelligent transportation systems.

Theory and algorithm for domain adaptation

Domain adaptation is an important subfield of transfer learning, it has been successfully applied in many applications of machine learning. Recently, significant theoretical and algorithmic advances have been achieved in domain adaptation. The theoretical analyses for domain adaptation are based on VC dimension and Rademacher complexity. There are also some covering number-based results, but most of these bounds are based on the results of Rademacher complexity, indirectly given by the relationship between covering number and Rademacher complexity. In this paper, we propose a theoretical analysis framework for domain adaptation, thus the error bound can be derived directly by covering number, which is an effective method for analyzing the generalization error in statistical learning theory. We derive generalization error bound for domain adaptation with a class of loss functions satisfying the assumptions. We also propose a mixup contrastive adversarial network for domain adaptation by introducing a mixup module for enhancing the alignment of the source and target domains during domain transfer, and a contrastive learning module for solving class-level alignment after domain transfer. Experimental results demonstrate the effectiveness of the proposed algorithm and the property of the theoretical results.

An Adaptive Time Filter Algorithm with Different Subdomain Time Steps for Super-Hydrophobic Proppants Based on the 3D Unsteady-State Triple-Porosity Stokes Model

An Adaptive Time Filter Algorithm with Different Subdomain Time Steps for Super-Hydrophobic Proppants Based on the 3D Unsteady-State Triple-Porosity Stokes Model