The existence of stochastic sampling phenomena in wastewater treatment processes (WWTPs) breaks the assumption that the existing control strategies use periodic data, and the operational constraints of equipment and the requirements for effluent water quality impose constraints on the system's input and output. These factors collectively increase the difficulty of achieving stable control of dissolved oxygen concentration (DOC). To solve these problems, a data-driven model predictive control (DDMPC) strategy is proposed to achieve stable control of constrained WWTPs with stochastic sampling intervals. First, a DDMPC framework is designed, which involves designing the objective function based on the mathematical expectation of the predicted output and considering system input and output constraints. In this framework, the problem of stochastic data acquisition caused by stochastic sampling can be solved, and the stable operation of the system can be ensured under constraints. Second, a data-driven multimodel prediction structure is constructed based on the stochastic characteristics of the sampling intervals. Specifically, fuzzy neural networks (FNNs) that match possible sampling intervals are established, thereby providing predictive outputs for the control process at the corresponding sampling instants. Third, a controller solving algorithm based on the generalized multiplier method is proposed, in which the constrained optimization problem within the model-predictive control (MPC) framework is reformulated by incorporating system constraints into the objective function as penalty functions to obtain the optimal control input that satisfies the constraints. Finally, the stability of the proposed DDMPC strategy is demonstrated, and its effectiveness is verified through the simulations on the benchmark simulation model No. 1 (BSM1). The results show that the proposed DDMPC strategy can achieve stable control of DOC in constrained WWTPs with stochastic sampling intervals.

Predefined time performance guaranteed sliding mode control for spraying robot based on Takagi-Sugeno fuzzy neural network

Exponential synchronization of T-S fuzzy complex-valued BAM neural networks with mixed time-varying delays via event-triggered control engineering and applications.

This article investigates the prescribed-time fault-tolerant cooperative control problem for second-order nonlinear heterogeneous multiagent systems (MASs) subject to multiple actuator faults. A novel prescribed-time stability criterion, independent of time-varying gain functions, is established to facilitate convergence analysis. Fuzzy neural networks (FNNs) are employed to approximate the unknown heterogeneous nonlinear dynamics, and a unified fault-tolerant control framework is constructed to simultaneously address both actuator bias and loss-of-effectiveness (LOE) faults. Based on a nonsingular sliding-mode approach, a distributed control protocol is designed to achieve prescribed-time tracking and containment consensus. Finally, simulation results for both single-leader and multileader scenarios validate the theoretical findings.

Comparison method-based finite-time stabilization of fuzzy complex-valued inertial neural networks with hybrid delays

BC-AFBTM: bubble-net communication optimization based adaptive fuzzy neural network with bidirectional long short-term memory for power quality improvement

The citrus diseases are affecting the fruit production worldwide thereby posing an economical burden. Major research is moving towards finding solutions using Artificial Intelligence (AI) and Image processing methods. Due to factors like illumination variations, leaf form, and disease symptoms, image data has intrinsic uncertainties that are typically difficult for traditional machine learning techniques to handle. In this paper, the interpretability of fuzzy logic is combined with the resilience of deep learning to propose a novel Fuzzy Convolutional Neural Network (Fuzzy-CNN) architecture for the automated diagnosis of citrus leaf diseases. The hybrid method uses a Convolutional Neural Network (CNN) to obtain complex features of citrus images, and a Fuzzy Inference System (FIS) to improve the classification results. The proposed approach encodes accurate data into fuzzy sets and applies linguistic concepts to determine the severity of a disease, which will contribute to the further development of the decision. In order to test and verify the proposed approach, several experiments were carried out, which proved that Fuzzy-CNN is more effective than regular CNN models with the approximate accuracy difference approximately 1.8, and especially in cases when the symptoms of disease are not clear. To strengthen experimental validation, the proposed method is evaluated on two independent datasets, including an external benchmark dataset, imbalance-aware evaluation metrics are employed to ensure robustness and generalizability. Experimental results demonstrate consistent and statistically significant improvements over existing neuro-fuzzy and machine learning approaches. This research contributes to early detection by collaborating the potential of fuzzy neural networks and offering a flexible solution for real-time disease detection in citrus crops.

Parkinson's disease (PD) is a progressive neurodegenerative disorder affecting millions of people worldwide. It severely impairs patients' mobility. For effective treatment strategies, it is essential to determine the severity of the disease at an early stage. In this study, an optimized feature-fusion framework is developed to detect the severity of Parkinson's disease using gait data via an Optimized Fuzzy Neural Network (OPtFuzNet). This feature-fusion framework can be categorized into three stages: pre-processing, feature extraction, and classification. In the pre-processing stage, a median filter is employed to reduce noise present in the gait images. During feature extraction, local and global features of the gait images are captured using SAE and IDCN, respectively. Subsequently, these features are fused and fed into the OPtFuzNet classifier, where parameter optimization is performed using IGWO-Lévy. The proposed model is evaluated on the GAIT-IT and GAIT-IST benchmark datasets using standard metrics, achieving superior performance with accuracies of 98.08% and 98.12%, respectively. Furthermore, a feature importance analysis is conducted to identify the most influential gait characteristics that contribute to determining the severity of the impairment, thereby enhancing the clinical interpretability of the model.The proposed model is evaluated on the GAIT-IT and GAIT-IST benchmark datasets using standard metrics, achieving superior performance with accuracies of 98.08% and 98.12%, respectively. Furthermore, a feature importance analysis is conducted to identify the most influential gait characteristics that contribute to determining the severity of the impairment, thereby enhancing the clinical interpretability of the model. These results validate the efficacy of the proposed feature fusion and optimization strategy in identifying discriminative gait patterns associated with the severity of Parkinson's disease. However, as the datasets are based on simulated gait patterns rather than real clinical data, further validation on real-world datasets is necessary for practical applicability.

This study proposes a fuzzy machine learning framework for optimizing antiepileptic drug selection using Quantitative Structure-Property Relationship (QSPR) modeling under pharmacological uncertainty. Feature relevance was assessed using Random Forest-based importance and SelectKBest with mutual information, and a feedforward neural network was trained with 5-fold cross-validation. Fuzzy membership functions were incorporated to model variability in clinical and experimental data. Compared with Multiple Linear Regression and conventional QSPR models, the proposed approach achieved a 24 percent reduction in RMSE. Predicted pharmacological attributes were further integrated into a Multi-Criteria Decision-Making framework using TOPSIS to rank drug candidates based on efficacy, safety, and cost. The resulting rankings showed 95 percent Spearman correlation with clinician evaluations, demonstrating the framework reliability for uncertainty-aware antiepileptic drug prioritization and custom MCDM libraries for TOPSIS based prioritization.

Multi-μ-stability and fixed-time multistability of switched fuzzy neural networks with discontinuous activation functions.

Fluid overload is common after neonatal congenital cardiac surgery (CCS) and is frequently managed with continuous furosemide infusions requiring iterative dose titration. An interpretable prediction model could support more consistent early postoperative dosing decisions. We hypothesized that a novel, interpretable machine learning approach could accurately predict furosemide dosing decisions in neonates following CCS. We identified term neonates admitted to the Pediatric Cardiothoracic ICU at a large academic children's hospital between 8/1/2014 and 3/1/2023 following CCS with cardiopulmonary bypass. Demographic and clinical data from the first 48 postoperative hours were used to train, validate, and test a Tropical Geometry-Based Fuzzy Neural Network Regressor (TGFNN-R) tasked with predicting furosemide infusion dose changes after CCS. The TGFNN-R was primed with clinician heuristics and provides transparent explanations behind predictions. A held-out internal validation/testing cohort was drawn from the same single-center population. Data from 506 neonates were extracted; 398 received a continuous furosemide infusion. Mean age at surgery was 6.2 (± 5.1) days; 67.3% were White. The most common surgeries were Stage I (Norwood) (25.1%) and arterial switch operation (18.6%). There were 783 furosemide dose increases and 224 dose decreases. Test set performance was R²=0.515, mean absolute error = 0.119mg/kg/hr, and false positive rate = 0.062. In this retrospective single-center cohort of neonates following CCS, an interpretable TGFNN-R model predicted and explained furosemide dose changes with good test performance. Next steps include external validation and nonclinical studies evaluating the model within clinical decision support and closed-loop paradigms to achieve prespecified fluid balance goals.

ABSTRACT In this study, an attention-mechanism-based convolutional long short-term memory (ACLSTM) model was integrated with a fuzzy neural network (FNN) model for the automatic multiclass classification of arrhythmias in electrocardiograms. The ACLSTM – FNN model uses convolutional neural network, long short-term memory networks, and an attention mechanism to extract local and regional temporal features. It also uses an FNN as a semantic classifier. Moreover, to prevent biases caused by data imbalance, the ACLSTM – FNN model employs class-balanced cross entropy as its loss function for training. In evaluation experiments, this model was used to classify arrhythmias in the MIT-BIH Arrhythmia Database into 17 categories. The ACLSTM – FNN model achieved accuracy, precision, recall, and F1 score values of 99.33%, 99.37%, 99.33%, and 99.34%, respectively, outperforming other models in identifying high-similarity classes.

A memristive fuzzy neural network with applications to classification task: A programmable circuit system.

Finite-time synchronization and anti-synchronization of fuzzy memristive competitive neural networks with reaction-diffusion terms

Synchronization of fractional-order delayed fuzzy memristive neural networks with unknown parameters and reaction-diffusion terms

Robust self-organizing fuzzy neural network with data immunity evaluation for industrial process modeling.

Prediction of combined cycle power plant electrical output power based on generalized dynamic interval type-2 fuzzy neural network

The problem of choosing a database architecture for storing and processing heterogeneous and dynamically changing data is considered. Modern information systems generate large amounts of structured, semi-structured and unstructured data from multiple sources, which makes the task of choosing an architecture multi-criteria and associated with high uncertainty. An intelligent decision support model based on a hybrid approach combining the methods of fuzzy logic and neural network analysis is proposed. Fuzzy logic is used to formalize expert knowledge and work with linguistic parameters such as “high load” or “low circuit flexibility”. It allows you to process blurred boundaries of criteria and build an IF–THEN rule system for architecture selection. The neural network component provides training based on historical data, identification of nonlinear patterns and adaptation to new conditions. Their integration is implemented in the form of a neuro-fuzzy model (for example, ANFIS), which combines the explicitness of the rules with the possibility of further training. The proposed architecture includes three levels: normalization and fuzzification of input data, a block of logical output with defuzzification and a corrective neural network subsystem. At the output, a probabilistic assessment of the priorities of architectural solutions is formed - relational, document–oriented, graph, column, or hybrid. Examples of rules, modeling results, and practical application scenarios are given: IP design, migration between databases, support for DevOps processes, and educational tasks. The practical significance of the research lies in reducing dependence on subjective expert assessments and increasing the reproducibility of architectural solutions. The prospects of expanding the rule base, applying deep network architectures, and integrating the model into engineering tools are emphasized.

This article focuses on the fixed-time synchronization problem for fractional-order fuzzy cellular neural networks (FOFCNNs) with information interactions and time-varying delays. To capture the complex dynamics of practical networks, nonlinear activation functions along with fuzzy AND and OR operators are incorporated into the master–slave systems. To achieve fixed-time synchronization despite these complexities, a novel adaptive multi-module controller is proposed. This controller integrates three functionally distinct components to accelerate the convergence rate, eliminate the effects of delays, and introduce negative feedback during communication, respectively. By employing fractional calculus tools, inequality techniques, and the proposed control law, sufficient criteria for the synchronization of the considered systems are rigorously established. Compared with existing synchronization works, this paper has significant advantages in model generality and controller design. Additionally, an explicit settling-time estimate is derived, which depends solely on control parameters and is independent of the initial conditions.

Abstract Buildings constructed prior to the implementation of seismic design standards or those built based on lower standards are susceptible to earthquake risks, resulting in substantial loss of life and property during an imminent earthquake. Although conventional rapid visual screening (RVS) methods have been extensively developed, both nationally and in the literature, they have limitations in accurately determining the vulnerability of buildings. Additionally, RVS methods developed on the basis of a single algorithm have limitations. Therefore, this study extends the existing body of work by integrating multiple AI algorithms, including fuzzy logic, machine learning, and neural networks, in the context of building damage data from the 2015 Gorkha earthquake, overcoming the limitations of previous studies by introducing an automated AI-based RVS methodology that enhances accuracy, transparency, and adaptability. The newly developed RVS method demonstrates an accuracy rate of 45.89% for testing in the three-class classification, while also delivering promising results in the two-class classification, with an accuracy rate of 60%, surpassing both conventional RVS methods and the baseline accuracy rate.