Gradient Descent Optimization Technique Research Articles

Research on passenger flow control at metro transfer stations based on real-time flow calculation of streamlines

PurposeThe volume of passenger traffic at metro transfer stations serves as a pivotal metric for the orchestration of crowd flow management. Given the intricacies of crowd dynamics within these stations and the recurrent instances of substantial passenger influxes, a methodology predicated on stochastic processes and the principle of user equilibrium is introduced to facilitate real-time traffic flow estimation within transfer station streamlines.Design/methodology/approachThe synthesis of stochastic process theory with streamline analysis engenders a probabilistic model of intra-station pedestrian traffic dynamics. Leveraging real-time passenger flow data procured from monitoring systems within the transfer station, a gradient descent optimization technique is employed to minimize the cost function, thereby deducing the dynamic distribution of categorized passenger flows. Subsequently, adhering to the tenets of user equilibrium, the Frank–Wolfe algorithm is implemented to allocate the intra-station categorized passenger flows across various streamlines, ascertaining the traffic volume for each.FindingsUtilizing the Xiaozhai Station of the Xi’an Metro as a case study, the Anylogic simulation software is engaged to emulate the intra-station crowd dynamics, thereby substantiating the efficacy of the proposed passenger flow estimation model. The derived solutions are instrumental in formulating a crowd control strategy for Xiaozhai Station during the peak interval from 17:30 to 18:00 on a designated day, yielding crowd management interventions that offer insights for the orchestration of passenger flow and operational governance within metro stations.Originality/valueThe construction of an estimation methodology for the real-time streamline traffic flow augments the model’s dataset, supplanting estimated values derived from surveys or historical datasets with real-time computed traffic data, thereby enhancing the precision and immediacy of crowd flow management within metro stations.

Deep Reinforcement Learning Technique for Traffic Metering in Connected Urban Street Networks

Proper metering can improve traffic operations in congested urban street networks. The available approaches either (a) use macroscopic fundamental diagrams to model traffic dynamics or (b) use numerical time–space discretization of the hydrodynamic traffic flow model with high computational requirements. Therefore, they either (a) do not represent traffic dynamics accurately or (b) are not suitable for online applications. This study introduces a deep reinforcement learning (DRL) methodology to capture traffic dynamics on a micro-level scale with the capability of capturing detailed traffic dynamics with low computational time. The DRL methodology employs two neural networks that map the location of connected vehicles in a network to traffic metering signal indications and estimate the objective function of the traffic metering problem. The outputs of the neural networks are used to construct a loss function, whose optimization provides the optimal parameters for the neural networks. Because of the complexity of the loss function, the gradient descent optimization technique with Monte Carlo simulation is used to optimize the loss function. The proposed methodology was tested on a simulated case study network in Vissim software with 20 intersections. The numerical results showed that the methodology increased throughput by 41.2% and 21.3% and reduced the total travel time of vehicles by 3.4% and 15.5% compared to a no-metering strategy. Comparing the computational time of the proposed methodology with one of the existing traffic metering approaches also showed the potential of the methodology for online applications.

GPINN with Neural Tangent Kernel Technique for Nonlinear Two Point Boundary Value Problems

Neural networks as differential equation solvers are a good choice of numerical technique because of their fast solutions and their nature in tackling some classical problems which traditional numerical solvers faced. In this article, we look at the famous gradient descent optimization technique, which trains the network by updating parameters which minimizes the loss function. We look at the theoretical part of gradient descent to understand why the network works great for some terms of the loss function and not so much for other terms. The loss function considered here is built in such a way that it incorporates the differential equation as well as the derivative of the differential equation. The fully connected feed-forward network is designed in such a way that, without training at boundary points, it automatically satisfies the boundary conditions. The neural tangent kernel for gradient enhanced physics informed neural networks is examined in this work, and we demonstrate how it may be used to generate a closed-form expression for the kernel function. We also provide numerical experiments demonstrating the effectiveness of the new approach for several two point boundary value problems. Our results suggest that the neural tangent kernel based approach can significantly improve the computational accuracy of the gradient enhanced physics informed neural network while reducing the computational cost of training these models.

A Novel Spider Monkey Optimization for Reliable Data Dissemination in VANETs Based on Machine Learning.

The growth in linked and autonomous vehicles has led to the emergence of vehicular ad hoc networks (VANETs) as a means to enhance road safety, traffic efficiency, and passenger comfort. However, VANETs face challenges in facilitating trustworthiness and high-quality services due to communication delays caused by traffic, dynamic topology changes, variable speeds, and other influencing factors. Hence, there is a need for a reliable data dissemination scheme capable of reducing communication delays among hops by identifying effective forwarder nodes. In this paper, we propose a novel, weighted, estimated, spider monkey-based, nature-inspired optimization (w-SMNO) method to generate a set of efficient relays. Additionally, we introduce a dynamic weight assignment and configuration model to enhance system accuracy using a neural network based on backpropagation with gradient descent optimization techniques to minimize errors in the machine learning model. The w-SMNO also incorporates a distinct algorithm for effective relay selection among multiple monkey spider groups. The simulation results demonstrate substantial improvements in w-SMNO, with a 35.7% increase in coverage, a 41.2% reduction in the end-to-end delay, a 36.4% improvement in the message delivery rate, and a 38.4% decrease in the collision rate compared to the state-of-the-art approaches.

Application And Optimization of Deep Reinforcement Learning in News Recommendation

In the era of online news platforms, the task of delivering personalized news recommendations to users has emerged as a critical challenge. This study delves into the utilization of Deep Q-Networks (DQN) within news recommendation systems, with a specific focus on the integration of loss functions and gradient descent optimization techniques. This combined approach aims to enhance the precision of estimating Q-values, ultimately resulting in more accurate and personalized article suggestions for users. The architectural design of the model involves the pairing of DQN with loss functions and gradient descent optimization, tailored for the domain of news recommendation. To validate this innovative approach, a comprehensive series of experiments has been executed, systematically benchmarking it against the conventional DQN framework. The empirical findings unequivocally demonstrate the superiority of the DQN fused with loss functions and gradient descent optimization across multiple performance metrics. These metrics encompass essential aspects such as click-through rates, user engagement duration, and overall satisfaction scores, affirming the effectiveness of the proposed approach. Furthermore, an extensive review of pertinent literature pertaining to the application of DQN in the realm of news recommendation is presented, providing readers with valuable contextual insights and a broader perspective. In summation, this paper underscores the compelling efficacy and untapped potential inherent in the fusion of loss functions and gradient descent optimization within DQN-based news recommendation systems.

Feedback control of millimeter scale pivot walkers using magnetic actuation

An external magnetic field can be used to remotely control small-scaled robots, making them promising candidates for diverse biomedical and engineering applications. In a previous study, we showed that our magnetically actuated millirobot is highly agile and can perform a variety of locomotive tasks such as pivot walking, tumbling, and tapping in a horizontal plane. In this study, we focus on controlling the locomotion outcomes of this millirobot in the pivot walking mode. A mathematical model of the system is developed and the kinematic model is derived. The role of the sweep and tilt angles in the robot’s motion is also investigated.We study two controllers to regulate the gait of the pivot walker. The first one is a proportional-geometric-based controller, which determines the correct pivot point that the millirobot should use. Then, it regulates the angular velocity proportionally based on the error between the center of the millirobot and the reference trajectory. The second controller is based on a gradient descent optimization technique, which expresses the control action as an optimization problem. These control algorithms enable the millirobot to generate a stable gait while tracking the desired trajectory. A low-cost high-performance magnetic actuator is built to validate the proposed controllers. We conduct a set of different experiments and simulation runs to establish the effectiveness of proposed controllers for different sweep and tilt angles in terms of tracking error. The two controllers exhibit an appropriate performance, but it is observed that the gradient descent-based controller yields faster convergence time, smaller tracking error, and fewer number of steps. Finally, we perform an extensive experimentally parametric analysis of the effect of sweep and tilt angles and step time on the tracking error. As we expect, the optimization-based controller outperforms the geometric-based one.

A novel range prediction model using gradient descent optimization and regression techniques

A novel range prediction model using gradient descent optimization and regression techniques

Gradient-descent iterative algorithm for solving exact and weighted least-squares solutions of rectangular linear systems

<abstract><p>Consider a linear system $ Ax = b $ where the coefficient matrix $ A $ is rectangular and of full-column rank. We propose an iterative algorithm for solving this linear system, based on gradient-descent optimization technique, aiming to produce a sequence of well-approximate least-squares solutions. Here, we consider least-squares solutions in a full generality, that is, we measure any related error through an arbitrary vector norm induced from weighted positive definite matrices $ W $. It turns out that when the system has a unique solution, the proposed algorithm produces approximated solutions converging to the unique solution. When the system is inconsistent, the sequence of residual norms converges to the weighted least-squares error. Our work includes the usual least-squares solution when $ W = I $. Numerical experiments are performed to validate the capability of the algorithm. Moreover, the performance of this algorithm is better than that of recent gradient-based iterative algorithms in both iteration numbers and computational time.</p></abstract>

A Machine Learning approach for collision avoidance and path planning of mobile robot under dense and cluttered environments

In this paper, a novel approach based on Machine Learning (ML) concept, i.e., the Adaptive Stochastic Gradient Descent Linear Regression (ASGDLR) algorithm, is developed to segregate an AMR's directional movement as right and left turn. Moreover, the developed algorithm is employed for path planning and navigational purposes. Here, real-time velocities of the right and left wheel and distance from obstacle data have been acquired by two Infrared (IR) sensors and one Ultrasonic (US) sensor positioned on the AMR. The weights of the proposed ASGDLR model are iteratively updated by applying Stochastic Gradient Descent (SGD) optimization technique by considering the difference between the actual velocity and model output velocity as an error signal. For the performance analysis of the proposed algorithm, three different performance indices, such as Mean Absolute Error (MAE), Mean Square Error (MSE), and Root Mean Square Error (RMSE), have been evaluated. The efficacy of the proposed algorithm is compared with six different AI-based regression algorithms for validation purposes. Moreover, the performance of the proposed algorithm is investigated for both obstacle avoidance and path planning in dense and cluttered environments. The simulation findings indicate that our proposed algorithm can accomplish tasks more efficiently than others while eliminating their shortcomings.

Time response optimal rational approximation: Improvement of time responses of MSBL based approximate fractional order derivative operators by using gradient descent optimization

Fractional calculus has become a popular topic in recent years. Its applications have been seen frequently in the analysis and solution of many engineering problems. Since, the exact realization of fractional order (FO) operators is not computationally feasible, several integer order approximations methods such as Oustaloup, Matsuda, Curve Fitting, modified stability boundary locus (MSBL) are utilized for frequency domain approximation of FO operators. These methods present advantages in frequency response approximation. However, real-world engineering applications require more accurate time responses to implement FO operators. Thus, any improvement in the approximate transfer function of FO systems will positively contribute to realization performance of FO system models in real-world applications. In this work, time responses of MSBL based approximate FO derivative models are improved by using gradient descent optimization (GDO) technique. Frequency domain approximation methods do not guarantee optimal approximation performance for the step response of FO elements. To deal with this weakness, author proposes a new GDO based rational approximation scheme that can considerably enhance the step response performance of MSBL method without deteriorating frequency response approximation performance. Results for the improved approximate FO model obtained by GDO (GDO-MSBL) are compared with the results of approximate FO models for popular FO approximation methods. The analog circuit realization is performed for the derivative operator, and the performance improvements for the proposed method are shown on the FO-PID controller and FO filter.

Intelligent breast cancer diagnostic system empowered by deep extreme gradient descent optimization.

Cancer is a manifestation of disorders caused by the changes in the body's cells that go far beyond healthy development as well as stabilization. Breast cancer is a common disease. According to the stats given by the World Health Organization (WHO), 7.8 million women are diagnosed with breast cancer. Breast cancer is the name of the malignant tumor which is normally developed by the cells in the breast. Machine learning (ML) approaches, on the other hand, provide a variety of probabilistic and statistical ways for intelligent systems to learn from prior experiences to recognize patterns in a dataset that can be used, in the future, for decision making. This endeavor aims to build a deep learning-based model for the prediction of breast cancer with a better accuracy. A novel deep extreme gradient descent optimization (DEGDO) has been developed for the breast cancer detection. The proposed model consists of two stages of training and validation. The training phase, in turn, consists of three major layers data acquisition layer, preprocessing layer, and application layer. The data acquisition layer takes the data and passes it to preprocessing layer. In the preprocessing layer, noise and missing values are converted to the normalized which is then fed to the application layer. In application layer, the model is trained with a deep extreme gradient descent optimization technique. The trained model is stored on the server. In the validation phase, it is imported to process the actual data to diagnose. This study has used Wisconsin Breast Cancer Diagnostic dataset to train and test the model. The results obtained by the proposed model outperform many other approaches by attaining 98.73 % accuracy, 99.60% specificity, 99.43% sensitivity, and 99.48% precision.

Cost Optimization in Neural Network using Whale Swarm Algorithm with Batched Gradient Descent Optimizer

Optimization algorithms are liable for sinking the losses and to give the most precise outcomes conceivable. Optimizers are utilized to modify the properties of neural network, for example, training rate and weights are used to reduce the losses. Optimization means a procedure of obtaining a global optimal solution for a given problem under given conditions. The real-world problems in the scientific fields, such as engineering design and economic planning, mostly are multimodal, high-dimensional, disconnected, and oscillated optimization problems. These complex problems cannot be solved well within reasonable time using traditional method based on gradient. Nature-inspired algorithms are becoming delightful in resolving mathematical optimization problems, like multiprocessor scheduling problem, vehicle routing and classification problems etc. In this manuscript, Whale Swarm Optimization algorithm on optimizing the neural networks, one of the meta-heuristic algorithms is applied to analysis of the cardiovascular disease dataset and compares the performance with Gradient Descent and RMSprop optimization techniques.

Analysis of Gradient Descent Optimization Techniques with Gated Recurrent Unit for Stock Price Prediction: A Case Study on Banking Sector of Nepal Stock Exchange

The stock price is the cost of purchasing a security or stock in a stock exchange. The stock price prediction has been the aim of investors since the beginning of the stock market. It is the act of forecasting the future price of a company's stock. Nowadays, deep learning techniques are widely used for identifying the stock trends from large amounts of past data. This research has experimented two big and robust commercial banks listed in the Nepal Stock Exchange (NEPSE) and compared stock price prediction performance of GRU with three widely used gradient descent optimization techniques: Momentum, RMSProp, and Adam. GRU with Adam is more accurate and consistent approach for predicting stock prices from the present study.

Model comparison of dark energy models using deep network

This work uses a combination of a variational auto-encoder and generative adversarial network to compare different dark energy models in light of observations, e.g., the distance modulus from type Ia supernovae. The network finds an analytical variational approximation to the true posterior of the latent parameters in the models, yielding consistent model comparison results with those derived by the standard Bayesian method, which suffers from a computationally expensive integral over the parameters in the product of the likelihood and the prior. The parallel computational nature of the network together with the stochastic gradient descent optimization technique leads to an efficient way to compare the physical models given a set of observations. The converged network also provides interpolation for a dataset, which is useful for data reconstruction.

Classification of malignant and benign tissue with logistic regression

Abstract Detection of breast cancer is the preliminary phase in cancer diagnosis. So, classifiers with higher accuracy are always desired. A classifier with high accuracy offers very less chance to wrongly classify a patient of cancer. This research investigates the performance of a modified and improved version of the hypothesis used in the logistic regression. Both gradient descent and advanced optimization techniques are used for the minimization of the cost function. A weighting factor of β is assigned in the hypothesis which is a sigmoid function. The dependency of the weighting factor to the number of features, the size of the dataset and the type of optimization technique used are observed. The accuracy of breast cancer detection is improved significantly by appropriately choosing the value of β, which, is a function of both the number of features and the type of optimization techniques used. The obtained results are promising by providing a significant increment in accuracy, sensitivity, and specificity.

Vehicle logo recognition using whitening transformation and deep learning

This paper presents a vehicle logo recognition using a deep convolutional neural network (CNN) method and whitening transformation technique to remove redundancy of adjacent image pixels. Backpropagation algorithm with stochastic gradient descent optimization technique has been deployed to train and obtain weight filters of the networks. Seven layers of our proposed CNN incorporating an input layer, five hidden layers and an output layer have been implemented to capture rich and discriminative information of vehicle logo images. Functioning as the output layer of the network, the softmax classifier is utilized to handle multiple classes of vehicle logo image. For a given vehicle logo image, the network provides the probability for each vehicle manufacturer to which the given logo image belongs. Unlike most of the common traditional methods that employ handcrafted visual features, our proposed method is able to automatically learn and extract high-level features for the classification task. The extracted features are discriminative sufficiently to perform well in various imaging conditions and complex scenes. We validate our proposed method by utilizing a public vehicle logo image dataset, which comprises 10,000 and 1500 vehicle logo images for training and validation objective, respectively. Experimental results based on our proposed method outperform other existing methods in terms of the computational cost and overall classification accuracy of 99.13%.

Estimation of Strain Elastography from Ultrasound Radio-Frequency Data by Utilizing Analytic Gradient of the Similarity Metric

Most strain imaging techniques follow a pipeline strategy: in the first step, tissue displacement is estimated from radio-frequency (RF) frames, and in the second step, a spatial derivative operation is applied. There are two main issues that arise from this framework. First, the gradient operation amplifies noise, and therefore, smoothing techniques have to be adopted. Second, strain estimation does not exploit the original RF data. It rather relies solely on the noisy displacement field. In this paper, a novel technique is proposed that utilizes both the displacement field and the RF frames to accurately obtain the strain estimates. The normalized cross correlation (NCC) metric between two corresponding windows around the samples of the pre- and post-compressed images is employed to generate a dissimilarity measurement. The derivative of NCC with respect to the strain is analytically derived using the chain rule. This allows an efficient minimization of the dissimilarity metric with respect to the strain using the gradient descent optimization technique. The effectiveness of the proposed method is investigated through simulation data, phantom experiments, and in vivo patient data. The experimental results show that exploiting the information in RF data significantly improves the strain estimates.

A Hybrid Algorithm for Short-Term Solar Power Prediction—Sunshine State Case Study

The growing rate of the integration of photovoltaic (PV) sites into the structure of power systems makes the task of solar power prediction more important in order to control the power quality and improve the reliability of system. In this paper, a hybrid forecasting algorithm is proposed for hour-ahead solar power prediction. A combination of gradient-descent optimization and meta-heuristic optimization approaches are designed in the structure of the presented model to take into account the prediction accuracy as well as the computational burden. At the first step, the gradient-descent optimization technique is employed to provide the initial parameters of a feedforward artificial neural network (ANN). At the next step, the meta-heuristic optimization model, called shuffled frog leaping algorithm (SFLA), is developed to search for the optimal set of parameters of ANN using the initial individuals found by the gradient-descent optimization. Then, the identified parameters by the customized SFLA will be employed by the ANN for short-term solar power prediction. The performance of the proposed forecasting algorithm is demonstrated on the solar power data of three simulated PV sites in Florida for 2006.

CONTROLLER PARAMETER OPTIMIZATION FOR AN ELECTRO-HYDRAULIC ACTUATOR SYSTEM BASED ON PARTICLE SWARM OPTIMIZATION TECHNIQUE

Motion control applications are widely applied in the industrial field with the assistance of electro-hydraulic actuator (EHA) system. EHA system is well known to be disclosed to the parameter variations, disturbances and uncertainties which are affected by the changes in the operating conditions such as friction, internal and external leakage. The complexity and nonlinear characteristic of the EHA system leads to a great challenge for controller development and system modelling. The performance of the utilized controller can be improved in order to achieve its best capability. In this paper, the basic knowledge in optimization of the proportional-integral-derivative (PID) controller through Gradient Descent (GD) and Particle Swarm Optimization (PSO) techniques are discussed. The PID parameters obtained through Ziegler-Nichols (ZN) tuning method has been optimized by using the GD method and is compared with PSO optimization technique via MATLAB/Simulink software. The findings show significant improvement in positioning tracking performance when the developed optimization technique is applied. Therefore, the issues degrading the EHA system performance have been reduced.

A Self-Adjusting Method for Real-Time Calculation of Thermal Loads in HVAC-R Applications

A significant step in the design of heating, ventilating, air conditioning, and refrigeration (HVAC-R) systems is to calculate room thermal loads. The heating/cooling loads encountered by the room often vary dynamically while the common practice in HVAC-R engineering is to calculate the loads for peak conditions and then select the refrigeration system accordingly. In this study, a self-adjusting method is proposed for real-time calculation of thermal loads. The method is based on the heat balance method (HBM) and a data-driven approach is followed. Live temperature measurements and a gradient descent optimization technique are incorporated in the model to adjust the calculations for higher accuracy. Using experimental results, it is shown that the proposed method can estimate the thermal loads with higher accuracy compared to using sheer physical properties of the room in the heat balance calculations, as is often done in design processes. Using the adjusted real-time load estimations in new and existing applications, the system performance can be optimized to provide thermal comfort while consuming less overall energy.