High Convergence Research Articles

Distributed optimal scheduling for virtual power plant with high penetration of renewable energy

The virtual power plants are aggregations of distributed generators, grid-connected devices in user side. The operation of virtual power plants affects the economic benefits and environmental benefits. However, due to the stochastic and fluctuating nature of power generation from renewable energy sources, the optimal scheduling problem for the virtual power plant containing high-dimensional random variables is difficult to solve. The conic power flow constraints are considered in the virtual power plant day-ahead optimal scheduling models to maximize the revenue by virtue of optimizing the flexible resources such as energy storage system and interruptible load. To quantify the uncertainties in the optimization problem, this paper proposes a network state-based power scenario reduction strategy for renewable energy generation, where typical scenarios are selected by the state of the grid voltage. The proposed day-ahead scheduling model is a mixed-integer, nonlinear, large-scale, stochastic optimization problem with high dimensional random variables, which is difficult to be solved directly by traditional centralized method. By temporally decoupling the charging and discharging model of energy storage systems, a distributed optimization algorithm is proposed based on multi-temporal power flow decoupling optimization model. The simulation results show that the proposed distributed optimization algorithm has high accuracy and good convergence, the virtual power plant can achieve day-ahead optimal scheduling and effectively promote renewable energy accommodation under the security and reliability of the grid.

Multi‐objective optimization design of concrete outriggers based on Genetic–HookeJeeves algorithm: Reducing lateral deflection, differential axial shortening, and construction cost of the structure

SummaryIn the context of tall concrete structures, it is crucial to not only control the lateral displacement of the building but also address the issue of differential axial shortenings in its vertical elements. Concrete outriggers, commonly resembling relatively stiff beams spanning one to two floors, connect the central core to exterior columns. The strategic placement and appropriate stiffness of these outriggers at different heights of the structure can significantly influence the overall behavior of the entire structure. This study focuses on optimizing the location, depth, and thickness of concrete outriggers, along with the dimensions of beams and columns, as well as the thickness of the core shear wall with the objective of minimizing construction costs and mitigating the occurrence of lateral displacement and differential axial shortenings within the structure. To achieve this, a combined approach of the Genetic–HookeJeeves algorithm has been employed. In this research, we have integrated HookeJeeves, a local search algorithm, with the genetic algorithm to create a hybrid approach that demonstrates high convergence performance. The structural modeling and analysis were conducted using ETABS finite element software, while a Euro‐International Concrete Committee model (CEB model) was utilized to assess the magnitude of differential axial shortenings, enabling us to approximate the long‐term behavior of concrete. The findings of this study highlight the significant impact of the location and stiffness of outriggers on mitigating both lateral displacement and differential axial shortenings within the structure. Optimal placement of an outrigger resulted in a 16% reduction in lateral displacement, and this value could reach up to 25% when the outrigger possessed the ideal stiffness. Additionally, such an arrangement led to a remarkable 36% decrease in the maximum differential axial shortening observed in the structure. These outcomes demonstrate that meeting the design requirements of the intended structure not only improves its performance but also reduces construction costs by 31%.

Cost-Effective Planning of Hybrid Energy Systems Using Improved Horse Herd Optimizer and Cloud Theory under Uncertainty

In this paper, an intelligent stochastic model is recommended for the optimization of a hybrid system that encompasses wind energy sources, battery storage, combined heat and power generation, and thermal energy storage (Wind/Battery/CHP/TES), with the inclusion of electric and thermal storages through the cloud theory model. The framework aims to minimize the costs of planning, such as construction, maintenance, operation, and environmental pollution costs, to determine the best configuration of the resources and storage units to ensure efficient electricity and heat supply simultaneously. A novel meta-heuristic optimization algorithm named improved horse herd optimizer (IHHO) is applied to find the decision variables. Rosenbrock’s direct rotational technique is applied to the conventional horse herd optimizer (HHO) to improve the algorithm’s performance against premature convergence in the optimization due to the complexity of the problem, and its capability is evaluated with particle swarm optimization (PSO) and manta ray foraging optimization (MRFO) methods. Also, the cloud theory-based stochastic model is recommended for solving problems with uncertainties of system generation and demand. The obtained results are evaluated in three simulation scenarios including (1) Wind/Battery, (2) Wind/Battery/CHP, and (3) Wind/Battery/CHP/TES systems to implement the proposed methodology and evaluate its effectiveness. The results show that scenario 3 is the best configuration to meet electrical and thermal loads, with the lowest planning cost (12.98% less than scenario 1). Also, the superiority of the IHHO is proven with more accurate answers and higher convergence rates in contrast to the conventional HHO, PSO, and MRFO. Moreover, the results show that when considering the cloud theory-based stochastic model, the costs of annual planning are increased for scenarios 1 to 3 by 4.00%, 4.20%, and 3.96%, respectively, compared to the deterministic model.

Solution of the Optimization Problem of Magnetotelluric Sounding in Quaternions by the Differential Evolution Method

The article discusses the application of quaternion Fourier transforms and quaternion algebra to transform Maxwell’s equations. This makes it possible to present the problem of magnetotelluric sensing (MTS) in a more convenient form for research. Studies of the inverse MTS problem for multi-layer regions are presented using the differential evolution method, which demonstrates high convergence. For single-layer regions, a new method for solving inverse problems based on minimizing the quadratic functional using conjugate optimization methods is considered. Numerical results obtained using special Python libraries are presented, with analysis and conclusions.

A geometrically nonlinear Hellinger–Reissner shell element for the postbuckling analysis of variable stiffness composite laminate structures

Variable stiffness (VS) composite laminates provide larger freedom to design thin-walled structures than constant stiffness (CS) composite laminates. They showed to allow the redistributing of stresses, improving buckling and post-buckling performance and, therefore, reducing material weight and costs. This work extends a recently developed mixed shell element, MISS-4C, to the postbuckling analysis of VS composite laminate structures. MISS-4C has a linear elastic closed-form solution for the stress interpolation of symmetric composite materials. Its stress field interpolation is obtained by the minimum number of parameters, making it an isostatic element. Moreover, its kinematic is only assumed along its contour, leading to an efficient evaluation of all operators obtained through analytical integration along the element contour. MISS-4C uses a corotational approach within a fast multi-modal Koiter algorithm to efficiently obtain the initial post-buckling response of VS composite laminate structures.First, the element performance is investigated by analysing a carbon fibre VS composite laminate plate subjected to compressive stresses. Numerical results obtained with MISS-4C are compared with those obtained with the MISS-4 element, showing good accuracy and a high convergence rate. Subsequently, the structural response of a glass fibre VS composite laminate girder of a short-length bridge is optimised through a multi-objective optimisation that exploits the robustness of the MISS-4C element and the efficiency of the multi-modal Koiter algorithm.

Heat and mass transfer analysis for bi-dimensional bioconvective MHD nanofluid with varying thermal traits

ABSTRACT Bioconvection phenomena offer practical applications in biomicrosystems, fuel cells, biosensors, biocomputing, electronic thermal management, and renewable energy systems, showcasing their relevance and potential impact across various engineering sectors. Consequently, this study delves into the examination of a bi-dimensional magnetohydrodynamic nanofluid flow model encompassing the intricate dynamics of gyrotactic microorganisms, subject to the effects of suction and thermal radiation. The primary objective of the investigation is to optimize the thermal energy transfer efficiency of the fluid while concurrently minimizing associated costs, with the secondary aim of diminishing the frictional resistance at the fluid-solid interface. The basis for energy and momentum equations incorporating radiation using Rosseland’s approximation is established by the Buongiorno nanofluid model. The foundational PDEs along with their BCs, are reformulated into ODEs utilizing similarity variables, then transformed into a set of first-order ODEs ensuring compatibility with MATLAB’s bvp4c solver, which offers high convergence rates and accuracy due to its efficient algorithms and precision handling. Heat transfer rate boosts up by increasing suction and radiation parameters, while it dwindles due to an increase in the magnetic field. Enhancing suction and the magnetic field lowers skin friction. Microbial concentration surges by increasing the magnetic field, while suction minimizes it. Intensifying the thermophoresis parameter increases the concentration of nanoparticles.

Optimal design of offshore jacket platform using enhanced colliding bodies optimization algorithm

Given the considerable volume of materials used in jacket platforms, structural optimization of these structures is always of interest. In the design optimization of offshore jacket platforms, the objective function is iteratively evaluated through a number of complex and time-consuming analyses making the optimization process computationally expensive. To reduce the computational costs, therefore, it is imperative to investigate efficient optimization algorithms with a high convergence rate to achieve optimal solutions for offshore jacket structures as a large-scale and complex problem. Accordingly, this research studies the application of a novel metaheuristic algorithm called Enhanced Colliding Bodies Optimization (ECBO) for the design optimization of a real jacket platform, SPD19A. The optimization constraints comprise stress and buckling in the members, horizontal displacements at the working point, and structural adequacy control of connections. The optimization results are subsequently compared to a design optimized by the Genetic Algorithm (GA), as an example, to evaluate the efficiency of the ECBO algorithm for the offshore jacket structure. The outcomes indicate that ECBO optimizes the jacket more effectively by 15%, while the optimization ratio of GA is 11%. Hence, the results confirm that ECBO has great and favorable efficiency and can potently escape from local optima to reach a better design for the jacket structure.

Convergence Analysis and Testing of Regional Effects of Green Finance on The Growth of Renewable Energy

ABSTRACT A paradigm shift in energy production and use is essential in the pursuit of carbon peaking and neutrality, and the expansion of renewable energy is a cornerstone of such a shift. This article first proposes an index that comprehensively measures renewable energy development in 30 provinces in China from 2010 to 2022. Then, the σ and β convergence models are applied to investigate the uniformity of renewable energy development. Furthermore, a spatial Durbin model is estimated to examine the spatial impact of green finance on renewable energy development. The results show that from 2010 to 2020, China has witnessed a significant increase in renewable energy use in its eastern, central and western regions, with the western region accelerating faster than its counterparts. During the same period, a remarkable trend of high convergence in renewable energy development has been observed, despite a slowdown in the spatial convergence rates between the central, eastern, western and national levels. The positive spatial spillover effect of renewable energy has increased significantly. The growth of renewable energy in the region and its surrounding areas is positively influenced by the growth of green finance.

Conjugate gradient‐based active noise cancellation

AbstractActive noise cancellation schemes should synthesize proper secondary perturbations without relevant prior information. Their performance depends strongly on their adaptive mechanisms, which, sometimes, do not fulfill the severe requirements of practical setups. This research demonstrates for the first time that the adaptive conjugate gradient optimization method can be used for active noise cancellation tasks. Additionally, a stochastic model is devised for estimating its steady‐state solution, even when the secondary path is not correctly estimated. Such a theoretical analysis demonstrates the asymptotic unbiasedness of the proposed scheme when the secondary plant is perfectly estimated. Simulations reveal that the devised method can attain better asymptotic performance and a higher convergence rate than the traditional FX‐LMS algorithm, at the expense of an increased computational burden. Furthermore, simulations show that the method outperforms the standard ones for nonstationary inputs (such as speech signals).

Quantum differential evolutionary algorithm with quantum-adaptive mutation strategy and population state evaluation framework for high-dimensional problems

Differential Evolution (DE) has been found to be inefficient and inaccurate in addressing high-dimensional complex problems. The Quantum-inspired Differential Evolution algorithm (QDE), endowed with quantum computing characteristics, efficiently manages high-dimensional problems but suffers from excessive mutation and poor convergence performance. Therefore, a new quantum differential evolutionary algorithm with quantum-adaptive mutation strategy and population state evaluation framework, namely PSEQADE is proposed. In PSEQADE, the quantum adaptive mutation strategy is employed to address the issue of excessive mutation in QDE, which adaptively reduces the degree of mutation, taking full advantage of the exceptional performance of quantum computing to enhance convergence accuracy. The quantum adaptive PSE framework is introduced to monitor the unstable mutation trends within the population, evaluate the population’s state, and intervene accordingly, thereby significantly improving the convergence performance and stability of the quantum differential evolution algorithm. 20 well-known functions from CEC2017 were selected for comparison with EPSDE, SADE, SHADE, JADE, CODE algorithms in dimensions of 500, 1000 and 3000. Additionally, comparisons were conducted with MLSHADE-SPA, SHADE-ILS, CCPSO2, NFDDE, DBO, and RIME algorithms in the dimension of 3000. Experimental results demonstrate that PSEQADE exhibits excellent convergence performance, high convergence accuracy, and exceptional stability in solving high-dimensional complex problems.

Effective multispike learning in a spiking neural network with a new temporal feedback backpropagation for breast cancer detection

This paper presents an effective learning multi-spike deep spiking neural network with temporal feedback backpropagation for breast cancer detection using contrast-enhanced MRI images. The learning in the spiking network is a new universal temporal feedback called Temporal_Feedback_SpikeProp (TF_SpikeProp) and Temporal_Feedback_ReSuMe (TF_ReSuMe). Thus, it can be implemented on all kinds of algorithms such as MultiSpikeProp and MultiReSuMe algorithms and is compatible with all temporal codings. The presented spiking network is a functional network with high accuracy and convergence in low epochs. The new algorithm explores the influence of all presynaptic neurons on the output error and reduces the role of inactive neurons. Therefore, the error propagation of the spiking network is such that it affects not only the spike time of each neuron in the output spike time, but also the refractory time and the presynaptic spikes of the neuron during the spiking time of the output neuron. The spiking network can correct and adjust the weights, spike delay, spike threshold, and also the time constant of the spike collection kernel. In order to diagnose cancer tissue, time–frequency features such as STFT and packet wavelet transform (WPT) have been used along with texture recognition features such as co-occurrence matrix. The presented network achieved an accuracy of 98.3% in 23 epochs in SpikeProp algorithm, 97.4% in ReSuMe with Duke MRI dataset, 98.64% on MNIST, and 96.1% on Iris. The results have shown that the training algorithm used in this study can achieve high accuracy in low epochs compared to traditional algorithms, while solving the challenge of getting stuck in local minima and poor convergence, as well as reducing the problem of time confusion of output spikes to an acceptable level and providing a fully functional network in practice and reality.

Dynamic Gaussian mutation beetle swarm optimization method for large-scale weapon target assignment problems

The weapon target assignment is a crucial issue for firepower resources optimization in modern warfare. Such a problem is complicated, multi-constrained, strongly nonlinear, NP-complete and the existing studies did not consider the suitability between different weapons and targets. In this paper, a novel weapon target assignment model is established that involves the weapon-target suitability and is closer to the real combat scenarios. Then, in view of that the conventional weapon target assignment methods are difficult to be applied in the large-scale problems efficiently, this work proposes a dynamic Gaussian mutation beetle swarm optimization algorithm with rule-based chaotic initialization. With the assistance of the dynamic parameter adjustment strategies and Gaussian mutation, the improved algorithm has fast convergence speed and high convergence accuracy, and it can solve the weapon target assignment problems with excellent optimization capabilities. Besides, the rule-based chaotic initialization strategy is embedded in this algorithm to generate high-quality population with better diversity. Finally, two comparative simulation cases of different initialization methods and algorithms for solving the large-scale weapon target assignment problems are designed. The results demonstrate that the proposed approach can provide more superior assignment schemes than its competitors with enhanced efficiency.

Convergence of the dimensional assessment of personality pathology (DAPP-BQ) and the five-factor personality inventory for the international classification of diseases 11th edition (FFiCD) in the context of the five-factor model and personality disorders

The current manuscript presents the convergence of the Dimensional Assessment of Personality Pathology (DAPP-BQ), using its short form the DAPP-90, and the Five-Factor Personality Inventory for International Classification of Diseases (ICD-11), the FFiCD, in the context of the five-factor personality model and the categorical approach of personality disorders (PDs). The current manuscript compares the predictive validity of both the FFiCD and the DAPP-90 regarding personality disorder scales and clusters. Results demonstrate a very high and meaningful convergence between the DAPP-90 and the FFiCD personality pathology models and a strong alignment with the FFM. The DAPP-90 and the FFiCD also present an almost identical predictive power of PDs. The DAPP-90 accounts for between 18% and 47%, and the FFiCD between 21% and 47% of PDs adjusted variance. It is concluded that both DAPP-90 and FFiCD questionnaires measure strongly similar pathological personality traits that could be described within the frame of the FFM. Additionally, both questionnaires predict a very similar percentage of the variance of personality disorders.

Numerical investigation on the convergence of self-consistent Schrödinger-Poisson equations in semiconductor device transport simulation

Semiconductor devices at the nanoscale with low-dimensional materials as channels exhibit quantum transport characteristics, thereby their electrical simulation relies on the self-consistent solution of the Schrödinger-Poisson equations. While the non-equilibrium Green’s function (NEGF) method is widely used for solving this quantum many-body problem, its high computational cost and convergence challenges with the Poisson equation significantly limit its applicability. In this study, we investigate the stability of the NEGF method coupled with various forms of the Poisson equation, encompassing linear, analytical nonlinear, and numerical nonlinear forms Our focus lies on simulating carbon nanotube field-effect transistors (CNTFETs) under two distinct doping scenarios: electrostatic doping and ion implantation doping. The numerical experiments reveal that nonlinear formulas outperform linear counterpart. The numerical one demonstrates superior stability, particularly evident under high bias and ion implantation doping conditions. Additionally, we investigate different approaches for presolving potential, leveraging solutions from the Laplace equation and a piecewise guessing method tailored to each doping mode. These methods effectively reduce the number of iterations required for convergence.

Analysis of Soil Slope Stability under Underground Coal Seam Mining Using Improved Radial Movement Optimization with Lévy Flight

Underground coal seam mining significantly reduces the stability of slopes, especially soil slopes, and an accurate evaluation of the stability of soil slopes under underground mining conditions is crucial for mining safety. In this study, the impact of coal seam mining is considered as the additional horizontal and vertical stresses acting on the slope, and an equation for calculating the safety factor of soil slopes under underground mining conditions is derived based on the rigorous Janbu method. Then, the Improved Radial Movement Optimization (IRMO) algorithm is introduced and combined with Lévy flight optimization to conduct global optimization searches, obtaining the critical sliding surface and corresponding safety factor of the soil slope under underground coal seam mining. Through comparisons with the numerical simulation results in three different case studies, the feasibility of applying the IRMO algorithm with Lévy flight to analyze the stability of soil slopes under underground mining is demonstrated. This ensures the accuracy and stability of the calculation results while maintaining a high convergence efficiency. Furthermore, the effects of the mining thickness and mining direction on slope stability are analyzed, and the results indicate that a smaller mining thickness and mining along the slope are advantageous for slope stability. The method proposed in this study provides valuable insights for preventing the slope instability hazards caused by underground coal seam mining.

High-Accuracy Photovoltaic Power Prediction under Varying Meteorological Conditions: Enhanced and Improved Beluga Whale Optimization Extreme Learning Machine

Accurate photovoltaic (PV) power prediction plays a crucial role in promoting energy structure transformation and reducing greenhouse gas emissions. This study aims to improve the accuracy of PV power generation prediction. Extreme learning machine (ELM) was used as the core model, and enhanced and improved beluga whale optimization (EIBWO) was proposed to optimize the internal parameters of ELM, thereby improving its prediction accuracy for PV power generation. Firstly, this study introduced the chaotic mapping strategy, sine dynamic adaptive factor, and disturbance strategy to beluga whale optimization, and EIBWO was proposed with high convergence accuracy and strong optimization ability. It was verified through standard testing functions that EIBWO performed better than comparative algorithms. Secondly, EIBWO was used to optimize the internal parameters of ELM and establish a PV power prediction model based on enhanced and improved beluga whale optimization algorithm–optimization extreme learning machine (EIBWO-ELM). Finally, the measured data of the PV output were used for verification, and the results show that the PV power prediction results of EIBWO-ELM were more accurate regardless of whether it was cloudy or sunny. The R2 of EIBWO-ELM exceeded 0.99, highlighting its efficient ability to adapt to PV power generation. The prediction accuracy of EIBWO-ELM is better than that of comparative models. Compared with existing models, EIBWO-ELM significantly improves the predictive reliability and economic benefits of PV power generation. This study not only provides a technological foundation for the optimization of intelligent energy systems but also contributes to the sustainable development of clean energy.

Multi-objective generalized normal distribution optimization: a novel algorithm for multi-objective problems

This study introduces the Multi-objective Generalized Normal Distribution Optimization (MOGNDO) algorithm, an advancement of the Generalized Normal Distribution Optimization (GNDO) algorithm, now adapted for multi-objective optimization tasks. The GNDO algorithm, previously known for its effectiveness in single-objective optimization, has been enhanced with two key features for multi-objective optimization. The first is the addition of an archival mechanism to store non-dominated Pareto optimal solutions, ensuring a detailed record of the best outcomes. The second enhancement is a new leader selection mechanism, designed to strategically identify and select the best solutions from the archive to guide the optimization process. This enhancement positions MOGNDO as a cutting-edge solution in multi-objective optimization, setting a new benchmark for evaluating its performance against leading algorithms in the field. The algorithm's effectiveness is rigorously tested across 35 varied case studies, encompassing both mathematical and engineering challenges, and benchmarked against prominent algorithms like MOPSO, MOGWO, MOHHO, MSSA, MOALO, MOMVO, and MOAOS. Utilizing metrics such as Generational Distance (GD), Inverted Generational Distance (IGD), and Maximum Spread (MS), the study underscores MOGNDO's ability to produce Pareto fronts of high quality, marked by exceptional precision and diversity. The results affirm MOGNDO's superior performance and versatility, not only in theoretical tests but also in addressing complex real-world engineering problems, showcasing its high convergence and coverage capabilities. The source codes of the MOGNDO algorithm are publicly available at https://nimakhodadadi.com/algorithms-%2B-codes.

Optimization of university management based on reptile search algorithm combined with short-duration memory neural network

Due to the lack of accuracy and difficulty to meet the actual requirements of the poor students' financial assistance in the management of smart colleges and universities. Therefore, a precise funding model for poor students is constructed on the basis of improved reptile search algorithm and short-term memory neural network. The performance of the algorithm is evaluated by test function, rank sum test and combinatorial model validation. In test function 5, the algorithm is 2.90E+01±6.04E-03, which is lower than the comparison algorithm, and begins to converge after about 10 iterations, and the convergence speed is significantly higher than that of the comparison algorithm. In the rank sum test, the experimental results of the comparison algorithm on most test functions are less than 5 %. In the combined model verification, the fitness result of the maximum convergence times was 0.2203 %, and the classification accuracy reached 98.7 %, which was better than the comparison model. The precise funding model of poor students proposed in this study has important application value in the management of smart colleges and universities, which can effectively improve the accuracy of poor students' funding and meet the actual needs. At the same time, the high accuracy and fast convergence of the model provide a new idea and method for smart university management.

A peridynamic-informed deep learning model for brittle damage prediction

Physics-informed Neural Network (PINN) has been introduced recently to predict and understand complex physical phenomena by directly incorporating feedback from governing equations. While PINNs offer remarkable capabilities, challenges arise when addressing discontinuities and heterogeneity, potentially compromising accuracy and reliability in prediction outcomes. In this study, a novel approach that combines the principles of peridynamic (PD) theory with PINN is presented to predict quasi-static damage and crack propagation in brittle materials. To achieve high prediction accuracy and convergence rate, the linearized PD governing equation is enforced in the PINN’s residual-based loss function. The proposed PD-INN is able to learn and capture complicated displacement patterns associated with different geometrical parameters, such as pre-crack position and length. The computational method presented in this study can be implemented to any PD constitutive model like Bond-Based PD or State-Based PD as well as any horizon factor. Several enhancements like cyclical annealing schedule and deformation gradient aware optimization technique are proposed to ensure the model would not get stuck in its trivial solution. The model’s performance assessment is conducted by monitoring the behavior of loss function throughout the training process. The results obtained through PD-INN demonstrate a satisfactory level of agreement in 2D and 3D when compared with the ground truth deformation from high-fidelity techniques like PD direct numerical method and Extended finite element method (X-FEM). Our results show the ability of the nonlocal PD-INN to predict damage and crack propagation accurately and efficiently.

Application of self-learning interval type-2 fuzzy neural network in PM2.5 concentration prediction

Abstract The change of PM2.5 concentration in air quality is nonlinear and difficult to predict. Therefore, a self-learning interval type-2 fuzzy neural network (SLIT2FNN) is proposed. SLIT2FNN has two parts: online structure learning and parameter learning. In structure learning, to improve the training accuracy and speed of the model, the Possibilistic Fuzzy C-Means (PFCM) algorithm is used to process the input data and obtain the number of initial rules before model training. The PFCM algorithm introduces the concept of possibility P to Fuzzy C-Means (FCM), allowing PFCM to overcome the shortcomings of FCM that cannot accurately cluster a large number of nonlinear problems. SLIT2FNN can establish an appropriate number of rules in the preparation stage, and then use the firing strength of the antecedents of the rules to judge whether to generate fuzzy rules for online self-learning, thereby optimizing its network structure. Then, the improved Levenberg–Marquardt (ILM) algorithm is used to modify the relevant parameters of SLIT2FNN. The ILM algorithm can address the challenge of numerous parameters in the Jacobian matrix and complex calculations and improve the calculation speed and adaptive ability of SLIT2FNN parameter learning. Finally, SLIT2FNN is applied to the prediction of air quality PM2.5 concentration, and the performance is compared with other models. The experiment proves that SLIT2FNN has a high prediction accuracy and fast convergence.