Trade-off Problem Research Articles

Balance Strategy Prediction Model Based on Power Load Data Mining

To meet the purpose of a multi-objective solution, a scheme is introduced into the optimal strategy-intensive training process. It supports a multi-task-driven computing model and has good reasoning accuracy, which makes the hydropower management system have efficient collaborative strategy prediction ability. Aiming at the trade-off problem between the cost-benefit and energy load of the hydropower control system, based on the neural network model to improve the stacked sparse denoising auto encoder second-order cone programming (SSDAE-SOCP) algorithm, a low-level backtracking depth uncertainty optimization scheduling model is proposed to obtain the optimal scheduling strategy. It intends to improve the poor accuracy of the global optimal results caused by the lack of priority sampling in the traditional strategy. The feature space without boundary conditions is set to solve the optimal solution for global/local organizations. Further, the flexibility is transferred according to the single-core function of task decomposition for improving the calculation accuracy of the whole solution data and the dynamic double-mutation balance. Simulation results have shown that the proposed algorithm enhances the training quality and performance by 9.5% and 15.2% compared with the traditional MOPSO and MOIMPA algorithms. Gradient descent and convergence speed, as well as the stability and security of intensive training, are better than the comparison methods. The training features verify the stability of the balanced prediction model, showing its importance in enhancing the anti-noise ability of the matrix. The research results can further enhance the multiple dispatching and command capabilities of the hydropower projects, providing technical support for the sustainable development of maximizing the benefits for the systems.

The effect of icon arrays and analogies in risk communication among adolescent children

The aim of the study was to measure the effect of icon arrays and analogies on the comprehension of risk information in adolescents aged 11–15 years. We tested whether icon arrays lead to higher accuracy in simple risk calculation tasks and in difficult tasks such as trade-off and Bayesian problems compared to the numerical format. We also measured whether analogies improved risk understanding. Icon arrays led to better understanding of risk information and more accurate risk comparisons. The effects varied according to the difficulty of the task and the risk literacy of the participants. Analogies were helpful for adolescents with high risk literacy.

Revisiting the Trade-Off Between Accuracy and Robustness via Weight Distribution of Filters.

Adversarial attacks have been proven to be potential threats to Deep Neural Networks (DNNs), and many methods are proposed to defend against adversarial attacks. However, while enhancing the robustness, the accuracy for clean examples will decline to a certain extent, implying a trade-off existed between the accuracy and adversarial robustness. In this paper, to meet the trade-off problem, we theoretically explore the underlying reason for the difference of the filters' weight distribution between standard-trained and robust-trained models and then argue that this is an intrinsic property for static neural networks, thus they are difficult to fundamentally improve the accuracy and adversarial robustness at the same time. Based on this analysis, we propose a sample-wise dynamic network architecture named Adversarial Weight-Varied Network (AW-Net), which focuses on dealing with clean and adversarial examples with a "divide and rule" weight strategy. The AW-Net adaptively adjusts the network's weights based on regulation signals generated by an adversarial router, which is directly influenced by the input sample. Benefiting from the dynamic network architecture, clean and adversarial examples can be processed with different network weights, which provides the potential to enhance both accuracy and adversarial robustness. A series of experiments demonstrate that our AW-Net is architecture-friendly to handle both clean and adversarial examples and can achieve better trade-off performance than state-of-the-art robust models.

Multi-module echo state network with variable skip length for chaotic time series prediction

Echo state networks (ESNs) have been extensively applied in time series prediction problems. However, the memory-nonlinearity trade-off problem severely limits the ability of ESNs to deal with chaotic time series prediction problems. In this study, a multi-module echo state network with variable skip length (MESN-VSL) is proposed to address this problem. First, the reservoir is divided into a nonlinear mapping module and multiple linear memory modules based on the idea of memory and nonlinearity separation. This idea can effectively balance the memory-nonlinearity trade-off problems. Second, a multi-module mechanism with skip length is put forward to model the characteristics of chaotic time series. The skip length and the number of linear memory modules of the MESN-VSL model are automatically determined based on the idea of phase-space reconstruction. Finally, the experimental results further demonstrate that the MESN-VSL model is superior to some existing models in chaotic time series prediction.

Achieving an appropriate polarization-breakdown synergy of multilayer films for dielectric energy storage through temperature gradient annealing

Large polarization and high breakdown strength are the key to achieving an idea energy storage density in dielectric capacitors, but unfortunately the trade-off problem between them is difficult to evade, particularly for those dielectrics with different crystalline degrees. Herein, we propose an appropriate polarization-breakdown synergistic strategy of dielectric multilayer films through temperature gradient annealing. The dielectric properties of a serials of (Pb, La)(Zr, Ti)O3/SrTiO3/(Pb, La)(Zr, Ti)O3 (PLZT/STO/PLZT) structures with different annealing temperature for each layer are compared. The optimum recoverable energy density of 57.9 J/cm3 is achieved at a high breakdown electric field of 5.78 MV/cm and a moderate maximum polarization of 22.5 μC/cm2. In addition to the role of suppressing the carriers transport of interlayer STO, the balance between polarization and breakdown strength in bottom and top PLZT layers with individual dielectric characteristics should be responsible for the enhanced energy storage performance (ESP). The results suggest that it is a feasible way to optimize the ESP of dielectric multilayer films through designing different stacking layer via temperature gradient annealing heat treatment.

ε-Constraint Procedures for Pareto Front Optimization of Large Size Discrete Time/Cost Trade-Off Problem

The discrete time/cost trade-off problem (DTCTP) optimizes the project duration and/or cost while considering the trade-off between activity durations and their direct costs. The complete and non-dominated time-cost profile over the set of feasible project durations is achieved within the framework of Pareto front problem. Despite the importance of Pareto front optimization in project and portfolio management, exact procedures have achieved very limited success in solving the problem for large size instances. This study develops exact procedures based on combinations of mixed-integer linear programming (MILP), ε-constraint method, network and problem reduction techniques, and present new bounding strategies to solve the Pareto problem for large size instances. This study also provides new large size benchmark problem instances aiming to represent the size of real-life projects for the DTCTP. The new instances, therefore, are generated to include up to 990 activities and nine execution modes. Computational experiments reveal that the procedures presented herein can remarkably outperform the state-of-the-art exact methods. The new exact procedures enabled obtaining the optimal Pareto front for instances with serial networks that include more than 200 activities for the first time.

PVC-based mixed-matrix membranes based on IL@AC/NH2-MIL-101 nanocomposites for improved CO2 separation performance

Mixed matrix membranes (MMMs), an important class of organic-inorganic nanocomposite membranes, were developed to overcome some of the limitations of purely polymeric membranes. In this study to improve the separation performance of polyvinyl chloride (PVC) membranes, mixed matrix membranes (MMMs) were prepared from incorporating choline prolinate based ionic liquid (IL) in a the coke/metal-organic framework (MOF) (NH2-MIL-101(Cr)) as a filler in polyvinyl chloride (PVC), which can be viewed as a potential solution to the trade-off problem with polymeric membranes because of the combination of the processing versatility of polymers and the high gas separation capability. Coke/MOF/PVC and IL@AC/MOF/PVC MMMs with different filler loadings of 5, 10, and 15 wt% were prepared using solution casting method and characterized using Fourier Transform Infrared Spectroscopy (FTIR), Thermogravimetric Analysis (TGA), Scanning Electron Microscopy (SEM) with Energy-Dispersive X-ray Spectroscopy (EDX) analyses, and Brunauer-Emmett-Teller (BET) surface area test. The porous structure of MMMs nanocomposites causes to which coke/MOF composite effectively accelerate gas diffusion in the PVC matrix. The permeability date was measured at 288.15, 298.15, 308.15 and 318.15 K and pressure up to 4 bar for CO2 and N2. According to the outcome, the addition of the IL([Cho][Pro]) filler, the permeability of the AC/MOF/PVC MMMs is increased compared to the pure PVC membrane. The MMMs have the highest gas separation efficiency and performance above Robson’s Upper Bound from 2008.

Simplified modeling study of the forward fuselage structure based on the overall structural rapid design software for civil aircraft

Abstract In order to solve the problems of poor weight estimation accuracy and difficult layout design trade-offs in the concept demonstration and overall program definition stage of civil aircraft, Aircraft Layout, a civil aircraft overall program structure rapid design software, is used to establish a numerical calculation-based layout program rapid design simulation platform. Based on the forward fuselage structure, this paper gives the results of rapid modeling and finite element simulation and analysis of the forward fuselage isotropic section of a certain type of aircraft, which provides a reference for the optimization design of the structural components of the forward fuselage section of a certain type of aircraft in the conceptual scheme design stage.

Sensorless robot collision detection based on fuzzy momentum observer

During physical human–robot interaction, unwanted collisions occur, which also threaten human safety. That is why different collision detection algorithms are proposed for safe collaboration between human and robots. The current detection methods are model-based methods without using additional sensors. However, the estimation of the collision force using a model-based method still has the problem of an unavoidable trade-off between the collision sensitivity and the reduction of the peaking value, and, in addition, the immunity to the model uncertainties. In this paper, an improved momentum observer using fuzzy system is proposed for robot collision detection. First, to avoid using the robot’s inverse inertia matrix, the generalized moment state equations based on the robot’s dynamics are deduced. Second, the extended linear momentum observer is built based on robot generalized momentum equations. Third, a fuzzy system is designed to select an appropriate bandwidth for the observer. In the design of this fuzzy system, the estimation error is considered as a fuzzy input, and the observer’s bandwidth is considered as its output. This fuzzy generalized momentum observer (FGMO) is compared with the existing generalized moment observers (GMOs) and has been proven its ability to overcome the mentioned above issues due to its ability to intelligently adjust the online filter parameters during the observation process. Finally, simulations are conducted with a 6-degree-of-freedom (6-DOF) industrial robot manipulator, and the obtained results illustrate the effectiveness of proposed approach in term of sensitivity and peaking value.

A Constraint Programming Approach for Discrete Time–Cost Tradeoff Problems in a Time-Constrained Activity Network

The discrete time-cost tradeoff problem (DTCTP) is a well-researched topic in the field of operations research. The majority of existing DTCTP models are based on traditional activity networks, which permit the execution of an activity as soon as all its predecessors have been completed. This assumption is reasonable, but it is important to note that there are always exceptions. The main work of this study was threefold. Firstly, we expanded the analysis of the DTCTP to encompass time-constrained activity networks (DTCTPTC), which encompassed three different types of time constraints. The first constraint was the time-window constraint, which limited the time interval during which an activity could be executed. The second constraint was the time-schedule constraint, which specified the times at which an activity could begin execution. The third constraint was the time-switch constraint, which required project activities to start at specific times and remain inactive during designated time periods. Secondly, a constraint programming (CP) model was developed for the purpose of solving the DTCTPTC. The model employed interval variables to define the activity and its potential time constraints, while CP expressions were utilized to ensure the feasibility of the solution. The objective was to identify the optimal execution mode for each activity, the optimal start times for time-scheduled activities, and the optimal work/rest patterns for time-switch activities, with the aim of minimizing the total cost of the project. Finally, the efficacy of the proposed CP model was validated through two case studies based on two illustrative projects of varying sizes. The outcomes were then compared against existing algorithms. The results demonstrated that time constraints were important factors affecting schedule optimization, and the proposed CP model had the ability to solve large-scale DTCTPTC.

An insight into physicochemical properties of hexagonal lyotropic liquid crystal templates from amphiphiles for mesoporous structural administration

Hexagonal lyotropic liquid crystals (HLLC) provide an ideal template for the fabrication of nanofiltration membranes without the flux-selectivity trade-off problem. Amphiphiles, the core components of the columnar micelles in the HLLC system, play a key role in regulating the nanostructure, their effects on dimensional stability and physicochemical properties of the system however have not been explored well. In this research, we compared the dimensional stability and physicochemical properties of HLLC templates prepared from amphiphiles with various head groups and tail lengths. The amphiphiles with charged head groups (dodecyl trimethylammonium bromide, DTAB, and hexadecyl trimethylammonium bromide, CTAB) enable more monomers to be introduced and a much smaller (∼ 1/2) pore size than that with uncharged ones (polyethylene glycol hexadecyl ether, Brij56). The tail length however mainly regulates the width of water channels with CTAB being 23.27 Å and DTAB about 17.06 Å when the head groups are identical. Although the HLLC systems with CTAB and DTAB present a little bit higher phase transition temperature, the dynamic mechanical stability is much better than that with Brij56 at room temperature, facilitating the reorientation process under an electric force with the order parameter reaching to about 0.9 and the retention of the aligned nanostructure with the absence of the force at room temperature. The study enables a better understanding of the effects of amphiphiles on HLLC templates and their tunability for advanced nanofiltration membranes.

Research on Privacy Issues in Smart Metering System: An Improved TCN-Based NILM Attack Method and Practical DRL-Based Rechargeable Battery Assisted Privacy Preserving Method

Smart meters, as a key component of Advanced Metering Infrastructure (AMI), collect fine-grained electricity consumption data for demand response in smart grids. While this data improves the grid’s accuracy, it also poses significant threats to users’ privacy. In this paper, we study the privacy issue in smart metering systems from both attacker and defender perspectives. First, we propose an improved Temporal Convolutional Network (TCN) based Non-Intrusive Load Monitoring (NILM) attack method, which infers electrical appliance usage from public load curves, addressing the gradient vanishing, gradient exploding, and other problems while improving attack accuracy. Second, we develop a rechargeable battery-assisted energy management system to hide load characteristics of electrical appliances by adding physical noise, thus resisting NILM attacks. To address the privacy-cost trade-off optimization problem, we propose a Practical Deep Reinforcement Learning-based Rechargeable Battery assist Privacy Preserving Method (PRoP) that learns optimal battery charging/discharging policies. We design a novel privacy measurement method and constraints to ensure the feasibility of system deployment and prove PRoP’s effectiveness in resisting NILM attacks. Comprehensive evaluations demonstrate that our improved TCN-based NILM method achieves an attack success rate of over 80% on various electrical appliances, improving attack performance (MAE, RMSE) by 20% compared to existing methods while reducing model training time. Moreover, our proposed PRoP achieves a better trade-off between privacy protection and electricity cost than existing battery-assisted methods, reducing costs by 5% and the attack success ratio to 36%, while increasing the MAE and RMAE obtained by NILM by 3 times. <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Note to Practitioners</i> —Smart grid, which can support bidirectional information transmission, has a series of advantages, such as high efficiency and high stability. However, it also brings a significant threat to users’ electricity privacy. Although encryption-based privacy protection methods have been deployed on terminal devices of smart grids to prevent privacy leaks, this method can often only defend against intrusive attacks and has little effect on non-intrusive attacks. To this end, this paper studies the privacy issues caused by non-intrusive attacks. Specifically, to better study the protection method, we first investigate the attack mechanism and design an improved TCN-based Non-Intrusive Load Monitoring method. Then, we propose a Practical Reinforcement learning-based rechargeable battery-assisted Privacy preserving method (PRoP) to defend against this attack physically. The most practical contribution of this paper is that, compared with existing battery-assisted privacy protection methods, we do not blindly pursue algorithm performance but fully consider the practical factors of deployment, such as limiting battery capacity and constraining battery charging and discharging behavior. This can guide practitioners to better apply this technology in practice.

Optimizing Three-Dimensional Trade-Off Problem of Time–Cost–Quality over Multi-Mode Projects with Generalized Logic

Clients typically tend to aim for reasonable prices, minimum possible makespans, and the best quality for the construction projects that they engage in. Evidently, weighing the available offers and coming up with an optimal decision can pose challenges for the decision makers. In this regard, the generation of a tool that helps decision makers strike a proper balance among the conflicting project objectives (i.e., time, cost, and quality) is imperative. To this end, this study proposes a method which assists in the selection of the best compromise choices among the options available for each of the project activities. In addition to the time and cost, the proposed method is designed to bring the quality aspect into the equation as well. To quantify the quality, a value referring to the weighted importance and performance of each activity is used. The proposed method is based on a modified multi-objective genetic algorithm (GA) that incorporates the domination concept for the selection of the best solutions out of the potential candidates. The GA-based method is capable of handling an unlimited number of precedence relationships for each activity, and above all, it is able to capture and unravel any type of logical relationship. This very feature significantly improves the practical relevance of this research, as the parallelization of activities is a common practice in real-life projects. Planners benefit from the various types of relationships (i.e., Start to Start, Start to Finish, Finish to Start, and Finish to Finish), and the concept of lag time frequently introduces parallelization into the network. Overlapped activities, in turn, help reduce the unwanted idle times and speed up the project significantly. Accordingly, in order to demonstrate the application and effectiveness of the proposed model, it has been used in the solution of four time–cost–quality (TCQ) trade-off problems, three of which are generated within the context of this paper. The practiced instances include a small benchmark TCQT problem with 18 activities taken from the literature in addition to more complex 29- and 63-activity TCQTPs produced herein based on benchmark time–cost trade-off problems. The performance of the presented approach is ultimately examined over a large-scale, real-case construction project with over four hundred activities and generalized logic in an unprecedented attempt to validate a model in the realm of TCQTPs. The successful results of the experiments reveal the effectiveness of the proposed model and corroborate the feasibility of its application by the planners amidst arduous decision-making processes.

On-board model predictive control for autonomous lane keeping with fuzzy preview distance: Design and experiment

This paper concerns with the development of computationally efficient lane keeping control method for the autonomous vehicles. To obtain autonomous lane keeping, model predictive control (MPC) scheme was intensively investigated in the previous studies owing to its inherent advantage of dealing with constrained multivariable systems. However, the tradeoff problem between the good tracking performance and the low computational complexity is inevitably raised in developing of MPC-based lane keeping technology. To alleviate the conflict between performance and cost, an on-board MPC with fuzzy preview distance is designed in this study. A linear system dynamics model is used in the MPC design to reduce the computational cost, and a fuzzy logic algorithm is developed to select an appropriate preview distance for enhancing the MPC performance. Further, hardware-in-the-loop test is adopted to explore the effectiveness and efficiency of the proposed control method. In comparison to the proportional-integral controller, the experimental results show that the MPC is more sensitive to the selective value of fixed preview distance. Since the significant impact of preview distance selection on the MPC-based lane keeping performance, the fuzzy logic algorithm is of the essence in terms of selecting the appropriate preview distance for MPC enhancement under different vehicle speed and road curvature. Eventually, experimental results validate that the proposed fuzzy preview distance algorithm can effectively improve the MPC-based lane keeping performance for autonomous vehicles subject to limited computational resource.

Optimizing time and cost in construction projects with a hybridized multi-verse optimizer and opposition-based learning

PurposeFor successful management of construction projects, a precise analysis of the balance between time and cost is imperative to attain the most effective results. The aim of this study is to present an innovative approach tailored to tackle the challenges posed by time-cost trade-off (TCTO) problems. This objective is achieved through the integration of the multi-verse optimizer (MVO) with opposition-based learning (OBL), thereby introducing a groundbreaking methodology in the field.Design/methodology/approachThe paper aims to develop a new hybrid meta-heuristic algorithm. This is achieved by integrating the MVO with OBL, thereby forming the iMVO algorithm. The integration enhances the optimization capabilities of the algorithm, notably in terms of exploration and exploitation. Consequently, this results in expedited convergence and yields more accurate solutions. The efficacy of the iMVO algorithm will be evaluated through its application to four different TCTO problems. These problems vary in scale – small, medium and large – and include real-life case studies that possess complex relationships.FindingsThe efficacy of the proposed methodology is evaluated by examining TCTO problems, encompassing 18, 29, 69 and 290 activities, respectively. Results indicate that the iMVO provides competitive solutions for TCTO problems in construction projects. It is observed that the algorithm surpasses previous algorithms in terms of both mean deviation percentage (MD) and average running time (ART).Originality/valueThis research represents a significant advancement in the field of meta-heuristic algorithms, particularly in their application to managing TCTO in construction projects. It is noteworthy for being among the few studies that integrate the MVO with OBL for the management of TCTO in construction projects characterized by complex relationships.

A new framework for project time–cost–environmental trade-off problem with hybrid Fermatean fuzzy–grey information

A new framework for project time–cost–environmental trade-off problem with hybrid Fermatean fuzzy–grey information

DCID: A divide and conquer approach to solving the trade-off problem between artifacts caused by enhancement procedure in image downscaling

Conventional research on image downscaling is conducted to improve the visual quality of the resultant downscaled image. However, there is an intractable problem, a trade-off relationship between artifacts such as aliasing and ringing, caused by enhancement procedure in image downscaling. To solve this problem, we propose a novel method that applies a divide-and-conquer approach for image downscaling (DCID). Specifically, the proposed DCID includes Weight-Net for dividing regions into enhancement first and artifact-least first regions and two generators that are optimized for divided regions to conquer the trade-off problem in the image downscaling task. The proposed method can generate a downscaled image without creating artifacts while preserving the perceptual quality of the input image. In objective and subjective evaluations, our experimental results show that the quality of the downscaled images generated by the proposed DCID is significantly better than benchmark methods.

Photovoltaic power prediction system based on multi-stage data processing strategy and improved optimizer

Building a reliable forecasting system can quantify future fluctuations in short-term photovoltaic output power, which is essential for optimizing grid configuration and reducing operating costs. However, most of the existing studies only use denoising technology to preprocess data, which results in the elimination of some key information. And the traditional optimizer cannot meet the parameter optimization requirements of the prediction system because of its limited search capacity and search space. Based on the above problems, a combined system based on multi-stage data processing strategy and improved optimizer is proposed, which solves the tradeoff problem between prediction accuracy and stability. Firstly, the multi-stage processing strategy effectively improves the signal-to-noise ratio and preserves more implicit information. Then, the optimal sub-model determination strategy extends the structural framework of model selection and improves the flexibility of the prediction system. Finally, three improved strategies are introduced to improve the optimization ability and convergence speed of the optimizer, which magnifies the advantages of the prediction system. An empirical study using Safi-Morocco data shows that the symmetric mean absolute percentage errors of three photovoltaic modules are 4.777%, 4.755% and 6.033%, respectively, which implies that the system can not only achieve accurate prediction of photovoltaic output power, but also help to balance supply and demand and improve the overall sustainability and stability of the grid-connected power system.

SIMULATION OF THE STRESSED-DEFORMED STATE OF THE ELASTIC DISCTOR RACK WITH A STIFFNESS REGULATOR

The paper involves the physical-mathematical modeling of the stress-strain state of an elastic stand of a developed disc harrow equipped with a stiffness regulator. The use of disc tools with individual attachment to elastic stands contributes to improving the quality and reducing the energy consumption of soil processing. This is due to the formation of oscillatory motion of the discs caused by the uneven resistance of the soil in the direction of the unit's movement. The goal of theoretical research is three-dimensional modeling of the stress-strain state of the elastic stand of the disc harrow with a stiffness regulator and substantiation of the range of its rational design parameters. To evaluate the interaction process between elastic working elements and the soil, a harmonic analysis was performed to determine the peak response of the system in a steady state to harmonic loads. In each step of the solution, all applied loads and base excitations have the same frequency, and the values are determined by the corresponding frequency curves. Using the SOLIDWORKS Simulation software package, the relevant physical-mathematical apparatus was compiled. On the basis of this approach, already at the stage of designing disk working bodies on elastic risers, it is possible to conclude about the possibility of them performing specific technological tasks, when by changing the parameters of the risers and conducting repeated studies, it is possible to obtain the corresponding dependencies in the form of regression equations. The results of numerical modeling provide visualization of the change in stress distribution over time in the stand. The dynamics of the change in maximum stress, located on the bend of the stand (R2) and the stiffness regulator (R1), were determined as the stress changes according to the law of damped oscillation with a specified natural frequency. In order to ensure the best working conditions of the elastic riser, it is necessary to ensure the greatest deformation of the riser at the place of attachment of the disc ΔL1 and, at the same time, the smallest value of the deformation of the tool frame ΔL2, by solving the trade-off problem in the Wolfram Cloud software package by maximizing the multiplicative function, the rational geometric dimensions of the diskator riser and constructive parameters of its placement in space.

An efficient and accurate multi-level cascaded recurrent network for stereo matching

With the advent of Transformer-based convolutional neural networks, stereo matching algorithms have achieved state-of-the-art accuracy in disparity estimation. Nevertheless, this method requires much model inference time, which is the main reason limiting its application in many vision tasks and robots. Facing the trade-off problem between accuracy and efficiency, this paper proposes an efficient and accurate multi-level cascaded recurrent network, LMCR-Stereo. To recover the detailed information of stereo images more accurately, we first design a multi-level network to update the difference values in a coarse-to-fine recurrent iterative manner. Then, we propose a new pair of slow-fast multi-stage superposition inference structures to accommodate the differences between different scene data. Besides, to ensure better disparity estimation accuracy with faster model inference speed, we introduce a pair of adaptive and lightweight group correlation layers to reduce the impact of erroneous rectification and significantly improve model inference speed. The experimental results show that the proposed approach achieves a competitive disparity estimation accuracy with a faster model inference speed than the current state-of-the-art methods. Notably, the model inference speed of the proposed approach is improved by 46.0% and 50.4% in the SceneFlow test set and Middlebury benchmark, respectively.