Convergence Problems Research Articles

Piggybacking on past problem for faster optimization in aluminum electrolysis process design

In process industry, it is a habitual phenomenon that the production efficiency is improved via replacing the obsolete devices with the advanced ones. To achieve an optimal performance, these devices are always required for heavily empirical adjustments, which is very time-consuming and inefficient. Though outdated, invaluable experiences for adjusting devices towards a more efficient industrial production have been accumulated by those replaced facilities. Thereby, new devices are expected to adapt to perform the industrial task without many modulations if the aforementioned experiences can be appropriately utilized. Inspired by the fact that evolutionary multitasking is capable of exploiting latent similarities and commonalities among multiple optimization tasks so as to improve the overall convergence of multi-task optimization, in this paper, we propose a novel framework to automatically search for the optimal settings for new devices based on the knowledge accumulated by the old. The framework, dubbed Piggybacking on Past Problem for Faster Optimization (PPPFO), is able to piggyback on the past optimization problem for a faster convergence of the targeted. By means of automatically transferring search experiences (i.e. genetic and cultural characteristics) from source task to the target, PPPFO can assist an engineering optimizer to improve its search exercises. PPPFO has been tested with a number of widely used benchmark functions and has been successfully adopted to an important real application, i.e., aluminum electrolysis process design. The remarkable results verify the efficacy and efficiency of the proposed framework.

Evaluation of soft soil foundation reinforcement effect of prefabricated building based on BP neural network

The use of traditional reinforcement methods in construction sites often causes problems such as pore water pressure, which can not effectively form a solid foundation. Aiming at this problem, the evaluation model of soft soil foundation reinforcement effect of prefabricated buildings is established based on BP neural network, combined with the geological characteristics of soft soil and the elements of foundation reinforcement; The L-M algorithm is used to optimize the slow convergence problem of BP neural network, and finally its evaluation effect is verified through practical application. The results show that the strengthening effect of 1550 kN⋅m/m2 is better than that of 2000 kN⋅m/m2 with the more times of tamping for marine and river facies, and there is a positive correlation between the times of strengthening and the effect. At the same time, similar qualitative conditions also show that the greater the burial depth, the worse the reinforcement effect. When the overlying soil layer is soft, the shallow buried soil layer can be reinforced by laying a cushion to improve the overall reinforcement effect. The laws reflected in the final model output data are the same as those reflected in the construction, and the accuracy of the proposed model is up to 87%, indicating that the model has superior performance in the reinforcement effect evaluation.

Automatic Modulation Recognition Based on a New Deep K-SVD Denoising Algorithm

Automatic modulation recognition has a wide range of applications in wireless communication. To solve the problem that the previous methods convert signal modulation recognition into image recognition, leading to increased time costs and information loss, an automatic modulation recognition approach consisting of the improved deep K-SVD denoising algorithm is suggested. First, the effectiveness of the model for random sine wave denoising is demonstrated by simulation. Second, the original I/Q signals are fed into the modified deep K-SVD model for denoising, skipping the complicated image processing steps. Finally, the noise-reduced signals are input into a multi-channel convolutional long short-term neural network to complete the modulation recognition. To solve the slow convergence problem of Iterative Shrinkage Thresholding Algorithms in sparse coding, the Fast Iterative Shrinkage Thresholding Algorithm is adopted to improve the computational efficiency and obtain a better denoising effect. The experiments show that the improved model has an average recognition accuracy of 91.26% at different SNRs from -2 dB to 18 dB, which is better than the state-of-the-art modulation recognition models.

Comfort Study of General Aviation Pilot Seats Based on Improved Particle Swam Algorithm (IPSO) and Support Vector Machine Regression (SVR)

Little work has been carried out to predict the comfort of aircraft seats, a component in close contact with the human body during travel. In order to more accurately predict the nonlinear and complex relationship between subjective and objective evaluations of comfort, this paper proposes a prediction method based on the Improved Particle Swarm Algorithm (IPSO) and optimized Support Vector Machine Regression (SVR). Focusing on the problems of the too-fast convergence and low accuracy of the traditional particle swarm algorithm (PSO), the improved particle swarm algorithm (IPSO) is obtained by linearly decreasing the dynamic adjustments of inertia weight ω, self-learning factor c1, and social factor c2; then, the penalty parameter C and kernel function parameter σ of SVR are optimized by the IPSO algorithm, and the comfort prediction of IPSO-SVR is established. The prediction accuracy of IPSO-SVR was 94.00%, the root mean square error RMSE was 0.37, the mean absolute value error MAE was 0.32, and the goodness of fit R2 was 0.92. The results show that the optimized IPSO-SVR prediction model can more accurately predict seat comfort under different angles and backrest tilt angles and can provide reference and research value for related industries. The results show that the optimized nonlinear prediction model of IPSO-SVR has higher accuracy, and its prediction method is feasible and generalizable, meaning it can provide a reliable basis for the prediction of seat comfort under different angles and backrest inclinations, as well as providing reference and research value for related industries.

Improving Li-ion battery health: Predicting remaining useful life using IWBOA-ELM algorithm

The prediction of Li-ion battery remaining useful life (RUL) is primarily used to prevent battery health damage caused by overcharging and discharging Li-ion batteries, which is critical for secondary battery life applications. To accurately predict the RUL of Li-ion batteries, a prediction method based on the improved black widow algorithm and extreme learning machine (IWBOA-ELM) is proposed in this paper. Because of the single method of black widow spider seeking, this paper employs the Archimedean spiral (AR) method to improve the search method of the BWOA algorithm and the global search capability of the BWOA algorithm; this is done to prevent the BWOA algorithm from being prone to failure. To avoid the problems of local optimal solutions and premature convergence that plague the BWOA algorithm, this paper employs the Golden-Sine strategy (Golden-SA) to update the positions of individual black widow spiders. Furthermore, for the cyclic trend of Li-ion batteries, this paper proposes an adaptive feedback factor to increase the convergence coefficient and control the early convergence rate while adapting to different battery capacity data. By comparing the outcomes of some CEC2017 test functions, the upgraded IBWOA performs well in determining the optimal. Finally, the IWBOA-ELM model is validated using NASA data, and the RUL prediction is performed under three scenarios: constant current discharge at different temperatures, constant current discharge at room temperature, and random current discharge at room temperature, with the results compared to the conventional method. The results reveal that the IWBOA-ELM method's RMSE error of RUL prediction is lower than that of other algorithms, indicating that IWBOA-ELM has strong generalization and robustness.

New robust approach for the globally optimal design of fired heaters

Fired heaters are important components of heat exchanger network (HEN) synthesis, especially in view of the existing procedures to simultaneously synthesize the HEN and obtain the basic design of heat exchangers. Existing basic design procedures for fired-heaters are either too simple, or difficult to solve when based on MINLP procedures. In addition, CFD approaches can hardly be used in design optimization and metaheuristics do not guarantee global optimality. We present two novel advances: a reformulation of models from previous optimization literature and we use Partial and Proxy Set Trimming followed by Smart Enumeration, a recently introduced robust technique. The proposed approach does neither needs initial solutions, nor does it have convergence problems and it also guarantees global optimality. The Set Trimming variant, namely Proxy Set Trimming, which was not part of the original Set Trimming version of Set Trimming is also used.

The Convergence Problem in Mean Field Games with Neumann Boundary Conditions

The Convergence Problem in Mean Field Games with Neumann Boundary Conditions

DQN-ALrM-Based Intelligent Handover Method for Satellite-Ground Integrated Network

The satellite-ground integrated network (SGIN), consisting of satellites and ground base stations, is regarded as the trend to solve the problem of global coverage. However, highly dynamic scenarios pose great challenges to users’ mobility management, especially handover decisions. With the rapid development of machine learning, reinforcement learning (RL), such as Q-learning and deep Q-network (DQN), has been applied to handover decision making owing to being good at solving dynamic multi-attribute optimization problems. However, Q-learning can be a waste of storage for large state spaces. Besides, the existing algorithms for training a neural network in a deep RL framework have problems of slow convergence or unstable convergence performance. Therefore, we propose a handover method exploiting the DQN framework of adaptive learning rate with momentum (DQN-ALrM), which can not only improve decision accuracy, but also improve learning efficiency. The customized DQN framework can solve problems with a large-size state space, and the proposed ALrM method can adjust the learning rate at any time according to the training error situation. Comprehensive simulations are conducted to validate the superiority of the proposed method in convergence speed, handover rate, call failure rate and multi-metric quality of service (QoS) over the alternative solutions in the literature.

A Comparative Study of Contrastive Learning-Based Few-Shot Unsupervised Algorithms for Efficient Deep Learning

The rapid development in computer vision algorithms, particularly in target detection algorithms based on deep neural networks, has led to significant advancements in various fields. However, recent research on target detection algorithms has shifted towards small sample scenarios with long-tailed distribution of categories, where achieving high accuracy target detection with limited data has become a crucial research topic. Deep neural networks often face several challenges when dealing with limited data, including convergence problems, overfitting, and poor generalization performance. In such scenarios, categories with little data are easily overwhelmed by the negative gradients generated by other categories during network training, which affects the final detection results. To address this issue, an unsupervised contrast learning algorithm has been proposed that can achieve accurate results on a small number of reference datasets without requiring a large number of datasets. One recently proposed contrast learning algorithm is PixelCL, which combines information from image depth in pretraining to improve results. The results of this algorithm show that it can obtain more accurate results through contrast learning with limited sample data. These findings highlight the potential of contrast learning algorithms in addressing negative gradient issues and improving the performance of deep neural networks in target detection tasks with limited data.

A two-stage approach for Bayesian joint models: reducing complexity while maintaining accuracy

Several joint models for longitudinal and survival data have been proposed in recent years. In particular, many authors have preferred to employ the Bayesian approach to model more complex structures, make dynamic predictions, or use model averaging. However, Markov chain Monte Carlo methods are computationally very demanding and may suffer convergence problems, especially for complex models with random effects, which is the case for most joint models. These issues can be overcome by estimating the parameters of each submodel separately, leading to a natural reduction in the complexity of the joint modelling, but often producing biased estimates. Hence, we propose a novel two-stage approach that uses the estimations from the longitudinal submodel to specify an informative prior distribution for the random effects when estimating them within the survival submodel. In addition, as a bias correction mechanism, we incorporate the longitudinal likelihood function in the second stage, where its fixed effects are set according to the estimation using only the longitudinal submodel. Based on simulation studies and real applications, we empirically compare our proposal with joint specification and standard two-stage approaches considering different types of longitudinal responses (continuous, count and binary) that share information with a Weibull proportional hazard model. The results show that our estimator is more accurate than its two-stage competitor and as good as jointly estimating all parameters. Moreover, the novel two-stage approach significantly reduces the computational time compared to the joint specification.

Hybrid Multi-Label Classification Model for Medical Applications Based on Adaptive Synthetic Data and Ensemble Learning.

In recent years, both machine learning and computer vision have seen growth in the use of multi-label categorization. SMOTE is now being utilized in existing research for data balance, and SMOTE does not consider that nearby examples may be from different classes when producing synthetic samples. As a result, there can be more class overlap and more noise. To avoid this problem, this work presented an innovative technique called Adaptive Synthetic Data-Based Multi-label Classification (ASDMLC). Adaptive Synthetic (ADASYN) sampling is a sampling strategy for learning from unbalanced data sets. ADASYN weights minority class instances by learning difficulty. For hard-to-learn minority class cases, synthetic data are created. Their numerical variables are normalized with the help of the Min-Max technique to standardize the magnitude of each variable's impact on the outcomes. The values of the attribute in this work are changed to a new range, from 0 to 1, using the normalization approach. To raise the accuracy of multi-label classification, Velocity-Equalized Particle Swarm Optimization (VPSO) is utilized for feature selection. In the proposed approach, to overcome the premature convergence problem, standard PSO has been improved by equalizing the velocity with each dimension of the problem. To expose the inherent label dependencies, the multi-label classification ensemble of Adaptive Neuro-Fuzzy Inference System (ANFIS), Probabilistic Neural Network (PNN), and Clustering-Based Decision tree methods will be processed based on an averaging method. The following criteria, including precision, recall, accuracy, and error rate, are used to assess performance. The suggested model's multi-label classification accuracy is 90.88%, better than previous techniques, which is PCT, HOMER, and ML-Forest is 65.57%, 70.66%, and 82.29%, respectively.

Optimizing a Multi-Layer Perceptron Based on an Improved Gray Wolf Algorithm to Identify Plant Diseases

Metaheuristic optimization algorithms play a crucial role in optimization problems. However, the traditional identification methods have the following problems: (1) difficulties in nonlinear data processing; (2) high error rates caused by local stagnation; and (3) low classification rates resulting from premature convergence. This paper proposed a variant based on the gray wolf optimization algorithm (GWO) with chaotic disturbance, candidate migration, and attacking mechanisms, naming it the enhanced gray wolf optimizer (EGWO), to solve the problem of premature convergence and local stagnation. The performance of the EGWO was tested on IEEE CEC 2014 benchmark functions, and the results of the EGWO were compared with the performance of three GWO variants, five traditional and popular algorithms, and six recent algorithms. In addition, EGWO optimized the weights and biases of a multi-layer perceptron (MLP) and proposed an EGWO-MLP disease identification model; the model was tested on IEEE CEC 2014 benchmark functions, and EGWO-MLP was verified by UCI dataset including Tic-Tac-Toe, Heart, XOR, and Balloon datasets. The experimental results demonstrate that the proposed EGWO-MLP model can effectively avoid local optimization problems and premature convergence and provide a quasi-optimal solution for the optimization problem.

Optimal planning and allocation of Plug-in Hybrid Electric Vehicles charging stations using a novel hybrid optimization technique.

India's expanding population has necessitated the development of alternate transportation methods with electric vehicles (EVs) being the most indigenous and need for the current scenario. The major hindrance is the undue influence on the power distribution system caused by incorrect charging station setup. Renewable Energy Sources (RES) have a lower environmental impact than the non-renewable sources of energy and due to which Plug-in Hybrid Electric Vehicles (PHEV) charging stations are installed in the highest-ranking buses to facilitate their effective placements. Based on meta-heuristic optimization, this study offers an effective PHEV charging stations allocation approach for RES applications. The primary objective of the developed system is to create a charging network at a reasonable cost while maintaining the operational features of the distribution network. These troublesare handled by applying meta-heuristic algorithms and optimum planning based on renewable energy systems to satisfy the outcomes of the variables. As a result, by adding charging station parameters, this research proposes to conceptualize the distribution of optimal charging stationsas multiple-objectives of the problem. Furthermore, the PHEV RES and charging station location problem is handled in this study by deploying a novel hybrid algorithm termed as Atom Search Woven Aquila Optimization Algorithm (AT-AQ) that includes the ideas of both Aquila Optimizer (AO) and Atom Search Optimization (ASO) Algorithms. In reality, Aquila Optimizer is a unique population-based optimization approach energized by Aquila's behaviour when seeking prey and it solves the problems of slow convergence and local optimum trapping. According to the findings of the experiments, the proposed model outperformed the other methods in terms of minimized cost function.

The Impact Modeling and Experimental Verification of a Launch Vehicle with Crushing-Type Landing Gear

In order to investigate the landing process of a vertical landing reusable vehicle, a dynamic model with a complex nonlinear dissipative element is established based on the discrete impulse step approach, which includes a three-dimensional multi-impact model considering friction and material compliance, and a multistage aluminum honeycomb theoretical model. The normal two-stiffness spring model is adopted in the foot–ground impact model, two motion patterns (stick and slip) are considered on the tangential plane and the structural changes caused by buffering behavior are included, and the energy conversion during the impact follows the law of conservation of energy. The state transition method is used to solve the dynamic stability convergence problem of the vehicle under the coupling effect of impact and buffering deformation in the primary impulse space. Landing experiments on a scaled physical reusable vehicle prototype are conducted to demonstrate that the theoretical results exhibit good agreement with the experimental data.

A Safety Warning Model Based on IAHA-SVM for Coal Mine Environment.

Coal is an important resource that is closely related to people's lives and plays an irreplaceable role. However, coal mine safety accidents occur from time to time in the process of working underground. Therefore, this paper proposes a coal mine environmental safety early warning model to detect abnormalities and ensure worker safety in a timely manner by assessing the underground climate environment. In this paper, support vector machine (SVM) parameters are optimized using an improved artificial hummingbird algorithm (IAHA), and its safety level is classified by combining various environmental parameters. To address the problems of insufficient global exploration capability and slow convergence of the artificial hummingbird algorithm during iterations, a strategy incorporating Tent chaos mapping and backward learning is used to initialize the population, a Levy flight strategy is introduced to improve the search capability during the guided foraging phase, and a simplex method is introduced to replace the worst value before the end of each iteration of the algorithm. The IAHA-SVM safety warning model is established using the improved algorithm to classify and predict the safety of the coal mine environment as one of four classes. Finally, the performance of the IAHA algorithm and the IAHA-SVM model are simulated separately. The simulation results show that the convergence speed and the search accuracy of the IAHA algorithm are improved and that the performance of the IAHA-SVM model is significantly improved.

Efficient adaptive reliability-based design optimization for geotechnical structures with multiple design parameters

Reliability-based design optimization (RBDO) aims to improve the reliability of geotechnical systems while minimizing the cost of design. However, the conventional RBDO method suffers from a large computational cost and intractable convergence problem due to the double-loop structure involving numerical optimization and repeated reliability analysis. To address this issue, this paper proposes an efficient adaptive RBDO framework for geotechnical structure with multiple design parameters. The proposed framework mainly includes two parts: a novel expanded inverse first-order reliability method (E-IFORM) for dealing with the inverse reliability problem with multiple design parameters, and an adaptive surrogate model to decouple the double-loop structure and improve the efficiency. The proposed methodology is applied to four geotechnical systems (i.e., a soil slope, an excavation pit, a strip footing and a retaining wall), and the results demonstrate the effectiveness and efficiency of the proposed framework.

A fast stereo matching network based on temporal attention and 2D convolution

We propose a fast stereo matching network based on temporal attention and 2D convolution (TANet). Due to the high similarity of the disparity between consecutive frames in an image sequence, we propose a temporal attention (TA) module that uses the disparity map of the previous frame to guide the disparity search range in the current frame, thus significantly improving the efficiency of disparity calculation in the cost volume module. Additionally, we propose a hierarchical cost construction and 2D convolution aggregation module that constructs a pyramid cost volume by fusing edge cues to establish detail constraints. This overcomes the problem of difficult convergence caused by information loss when replacing 3D convolution with 2D convolution. Experimental results show that the TA module effectively optimizes the cost volume and, together with 2D convolution, improves the computational speed. Compared with state-of-the-art algorithms, TANet achieves a speedup of nearly 4x, with a running time of 0.061s, and reduces the parameter count by nearly half while decreasing accuracy by 1.1%. Code is available at https://github.com/Y0uchenZ/TANet.

Some Convergence Problems Regarding the Fractional Schrödinger Propagator on Noncompact Manifolds

Some Convergence Problems Regarding the Fractional Schrödinger Propagator on Noncompact Manifolds

Modelling of photovoltaic cells using Boole’s rule-based multi-objective genetic algorithm implemented in Indoor hardware setup

Triple Diode Solar Cell Module (TDSCM) circuit with nine parameters for various environmental circumstances represents the behavior and practical performance of solar cell.The precise extraction of photovoltaic (PV) module parameters is essential for optimising the energy conversion efficiency of PV systems. Usually the equations describing solar panels are implicit in nature, and parameter extraction has been very complicated. The solar cell is mathematically modelled with nonlinear I-V (Current – Voltage) characteristics behavior, and it cannot be directly determined from the PV’s datasheet due to the lack of data offered by the PV manufacturers. On the basis of the technical datasheet of the photovoltaic module (PV), only four equations can be obtained in single diode, double diode, and triple diode parameters. To be implemented with fifth equation, many researchers have been done with multiple approximations and it becomes with low accuracy, complexity of computation, convergence problem. To resolve these issues, a new multi-objective optimization (GA) genetic algorithm method is prescribed to frame the fifth equation using the Boole rules implemented with the curved area concept. The proposed Boole’s rule based model offers superior non-linearity performance and high precision modelling, and the error shows a significant reduction when compared to the single and double diode approaches used in the existing approach. The effectiveness of the proposed I-V curve characteristics efficiency was improved by the implementation of the proposed Boole’s rule with RMSE error 0.000034.

Multi-objective sustainable supply chain network optimization based on chaotic particle-Ant colony algorithm.

For the optimal design of the sustainable supply chain network, considering the comprehensiveness of the problem factors, considering the three aspects of economy, environment and society, the goal is to minimize the establishment cost, minimize the emission of environ-mental pollution and maximize the number of labor. A mixed integer programming model is established to maximize the efficiency of the supply chain network. The innovation of this paper, first, is to consider the impact of economic, environmental and social benefits in a continuous supply chain, where the environmental benefits not only consider carbon emissions but also include the emissions of plant wastewater, waste and solid waste as influencing factors. Second, a multi-objective fuzzy affiliation function is constructed to measure the quality of the model solution in terms of the overall satisfaction value. Finally, the chaotic particle ant colony algorithm is proposed, and the problem of premature convergence in the operation of the particle swarm algorithm is solved. Experimental results show that the PSCACO algorithm proposed in this paper is compared with MOPSO, CACO and NSGA-II algorithms, and the convergence effect of the algorithm is concluded to be more effective to verify the effectiveness and feasibility of chaotic particle ant colony algorithm for solving multi-objective functions, which proposes a new feasible solution for the supply chain management.