Sigmoid Function Research Articles

Aperiodic EEG Predicts Variability of Visual Temporal Processing.

The human brain exhibits both oscillatory and aperiodic, or 1/f, activity. Although a large body of research has focused on the relationship between brain rhythms and sensory processes, aperiodic activity has often been overlooked as functionally irrelevant. Prompted by recent findings linking aperiodic activity to the balance between neural excitation and inhibition, we investigated its effects on the temporal resolution of perception. We recorded electroencephalography (EEG) from participants (both sexes) during the resting state and a task in which they detected the presence of two flashes separated by variable interstimulus intervals. Two-flash discrimination accuracy typically follows a sigmoid function whose steepness reflects perceptual variability or inconsistent integration/segregation of the stimuli. We found that individual differences in the steepness of the psychometric function correlated with EEG aperiodic exponents over posterior scalp sites. In other words, participants with flatter EEG spectra (i.e., greater neural excitation) exhibited increased sensory noise, resulting in shallower psychometric curves. Our finding suggests that aperiodic EEG is linked to sensory integration processes usually attributed to the rhythmic inhibition of neural oscillations. Overall, this correspondence between aperiodic neural excitation and behavioral measures of sensory noise provides a more comprehensive explanation of the relationship between brain activity and sensory integration and represents an important extension to theories of how the brain samples sensory input over time.

AONet: Attention network with optional activation for unsupervised video anomaly detection

AbstractAnomaly detection in video surveillance is crucial but challenging due to the rarity of irregular events and ambiguity of defining anomalies. We propose a method called AONet that utilizes a spatiotemporal module to extract spatiotemporal features efficiently, as well as a residual autoencoder equipped with an attention network for effective future frame prediction in video anomaly detection. AONet utilizes a novel activation function called OptAF that combines the strengths of the ReLU, leaky ReLU, and sigmoid functions. Furthermore, the proposed method employs a combination of robust loss functions to address various aspects of prediction errors and enhance training effectiveness. The performance of the proposed method is evaluated on three widely used benchmark datasets. The results indicate that the proposed method outperforms existing state‐of‐the‐art methods and demonstrates comparable performance, achieving area under the curve values of 97.0%, 86.9%, and 73.8% on the UCSD Ped2, CUHK Avenue, and ShanghaiTech Campus datasets, respectively. Additionally, the high speed of the proposed method enables its application to real‐time tasks.

Physics-informed unsupervised domain adaptation framework for cross-machine bearing fault diagnosis

Varying components and operating conditions in industrial machines lead to different distribution characteristics and fault states of monitoring data for critical rotating machinery. Adequate and accurate labelling of monitoring data is challenging to achieve. Existing data-driven unsupervised domain adaptation methods exhibit poor interpretability and low confidence in the target domain’s pseudo-labelling. Therefore, a physics-informed unsupervised domain adaptation framework is proposed. Data and physical information are combined to enhance the reliability and interpretability of the model while ensuring its generalization capability. First, physical information is extracted by merging the knowledge of the mechanical component fault characteristics with indicators like the signal spectrum’s energy value. Subsequently, the physical label and loss are created. The physical loss is then embedded in the loss function of the source domain model. Next, by aligning the physical labels and the clustering results of the spectral clustering, the target domain samples’ hard pseudo-labels are obtained. The probability values of the sample belonging to each category as determined by the Wasserstein criterion comprise the soft pseudo-label of the sample. Furthermore, the confidence of hard pseudo-labels is continually increased by iteratively updating the probability values of soft pseudo-labels using the multi-label as a guide and the inverse improved sigmoid function as a gradient. Finally, unsupervised cross-machine diagnostic models with high robustness can be obtained. Three datasets are employed in the experimental section, and five sub-datasets are created and crossed for cross-machine experiments to confirm the suggested approach’s efficacy and superiority.

Dynamic Surface Tracking Control for Nonlinear Constraints

Abstract An adaptive control scheme based on dynamic surface control (DSC), and intelligent control techniques is developed for a 3-DOF helicopter with unknown disturbance and asymmetric saturation. The reconstructed dynamic neural network (NN) is established to compensate for nonlinear terms in real-time. A designed smooth nonlinear sigmoid function is used to model saturation nonlinearity. The combination of dynamic surface control with adaptive intelligent control techniques resolves and improves the tracking and robust performance of the system. The stability and ultimately bound of all signals of the controlled system are demonstrated by theoretical analysis. Numerical results prove the feasibility of the presented control scheme.

Investigation of the effect of the flow rate of coolant on the kinetics of carbon dioxide hydrate growth using sigmoidal growth model

Relaxation time plays a crucial role in studying reaction kinetics and other processes. This study addresses the importance of relaxation time and suggests methods to minimize it. In this article, we adjust the relaxation time by extracting heat from the hydrate formation process through variations in the flow rate of the cooling fluid. It has been demonstrated that heat transfer plays a dominant role in this process. The kinetics of carbon dioxide hydrate growth were evaluated by increasing the flow rate of cooling fluid using modified sigmoidal growth curve equations, including Logistics and Gompertz. It was conducted a nonlinear least-squares regression analysis to fit sigmoidal functions to the cumulative formation curves of hydrate generated from gas uptake in the static reactor over time. The percentage of error between Areal and Amodel shows that the Gompertz model with Q = 14.59 Lit/min at T = 276 K and Q = 20.25 Lit/min at T = 277 K is the best model for predicting the maximum consumption capacity of CO2. The results also showed that increasing the cooling fluid’s flow rate reduces relaxation time at any temperature. Moreover, increasing the flow rate of the cooling fluid decreased the average relaxation time by 2 % to 60 % at T = 276 K and by 7 % to 22 % at T = 277 K compared to the lowest investigated flow rate. Additionally, reducing the experimental temperature while keeping the flow rate of the cooling fluid constant led to a reduction in the relaxation time.

Multi-physics numerical simulation study on thermo-sensitive gel delivery for a local post-tumor surgery treatment

Numerous studies in the literature have proposed the use of thermo-responsive hydrogels for filling cavities after tumor resection. However, optimizing the injection process is challenging due to the complex interplay of various multi-physics phenomena, such as the coupling of flow and heat transfer, multi-phase interactions, and phase-change dynamics. Therefore, gaining a fundamental understanding of these processes is crucial. In this study, we introduce a thermo-sensitive hydrogel formulated with poloxamer 407 and Gellan gum as a promising filling agent, offering an ideal phase-transition temperature along with suitable elastic and viscous modulus properties.We performed multi-physics simulations to predict the flow and temperature distributions during hydrogel injection. The results suggested that the hydrogel should be kept at 4 °C and injected within 90 s to avoid reaching the transition temperature. Cavity filling simulations indicated a symmetric distribution of the hydrogel, with minimal influence from the syringe's position.The temperature gradient at the cavity edge delays gelation during injection, which is essential to guarantee its administration as a liquid. The hydrogel's viscosity follows a sigmoidal function relative to temperature, taking five minutes to reach its maximum value. In summary, the multi-physics simulations carried out in this study confirm the potential of thermo-responsive hydrogels for use in post-tumor surgery treatment and define the conditions for a proper administration. Furthermore, the proposed model can be widely applied to other thermo-responsive hydrogels or under different conditions.

MRI-Based Deep Learning for Differentiating Between Bipolar and Major Depressive Disorders

Mood disorders, particularly bipolar disorder (BD) and major depressive disorder (MDD), manifest changes in brain structure that can be detected using structural magnetic resonance imaging (MRI). Although structural MRI is a promising diagnostic tool, prevailing diagnostic criteria for BD and MDD are predominantly subjective, sometimes leading to misdiagnosis. This challenge is compounded by a limited understanding of the underlying causes of these disorders. In response, we present SE-ResNet, a Residual Network (ResNet)-based framework designed to discriminate between BD, MDD, and healthy controls (HC) using structural MRI data. Our approach extends the traditional Squeeze-and-Excitation (SE) layer by incorporating a dedicated branch for spatial attention map generation, equipped with soft-pooling, a 7 × 7 convolution, and a sigmoid function, intended to detect complex spatial patterns. The fusion of channel and spatial attention maps through element-wise addition aims to enhance the model's ability to discriminate features. Unlike conventional methods that use max-pooling for downsampling, our methodology employs soft-pooling, which aims to preserve a richer representation of input features and reduce data loss. When evaluated on a proprietary dataset comprising 303 subjects, the SE-ResNet achieved an accuracy of 85.8%, a recall of 85.7%, a precision of 85.9%, and an F1 score of 85.8%. These performance metrics suggest that the SE-ResNet framework has potential as a tool for detecting psychiatric disorders using structural MRI data.

Vulnerability of AND/OR logic networks under cascading failures

Cascading failures in AND/OR logic networks present a critical challenge in domains such as digital circuits and communication systems where reliability and resilience are paramount. This work investigates the vulnerability of AND/OR logic networks under cascading failures by examining the propagation mechanisms, failure modes, and robustness of such systems. In detail, we propose a hybrid fault propagation model based on the basic principles of AND/OR logic operations. Then, by utilizing both analytical and simulation approaches, we explore how initial localized disruptions can escalate through interconnected logical elements and lead to widespread system failures. Notably, it is shown that a Sigmoid function can be developed to quantify the vulnerability of the complex logic network. Moreover, the phase transition of the failures against the logic parameters is depicted and it is illustrated that the phase transition point is independent of the AND links. This study contributes to the understanding of cascading dynamics in logical networks and could facilitate design of feasible strategies for mitigating these vulnerabilities and recovering from substantial failures.

Q-Deformed and delta-parametrized A-generalized logistic function induced Banach space valued multivariate multi layer neural network approximations

Abstract. Here we research the multivariate quantitative approximation of Banach space valued continuous multivariate functions on a box or RN, N ∈ N, by the multivariate normalized, quasi-interpolation, Kantorovich type and quadrature type neural network operators. We investigate also the case of approximation by iterated multilayer neural network operators of the last four types. These approximations are achieved by establishing multidimensional Jackson type inequalities involving the multivariate modulus of continuity of the engaged function or its partial derivatives. Our multivariate operators are defined by using a multidimensional density function induced by a q-deformed and λ-parametrized A-generalized logistic function, which is a sigmoid function. The approximations are pointwise and uniform. The related feed-forward neural network are with one or multi hidden layers. Mathematics Subject Classification (2010): 41A17, 41A25, 41A30, 41A36. Keywords: Multi-layer approximation, q-deformed and λ-parametrized A- generalized logistic function, multivariate neural network approximation, quasi-interpolation operator, Kantorovich type operator, quadrature type operator, multivariate modulus of continuity, iterated approximation.

Predictive framework for remaining useful life of roller bearings: Utilizing fractional generalized Pareto degradation model in performance evaluation

Research on predictive frameworks for the Remaining Useful Life (RUL) of mechanical systems is limited. This study proposes a comprehensive RUL prediction framework for roller bearings, integrating early failure assessment, adaptive failure threshold determination, and an adaptive drift fractional-order Pareto degradation model. To tackle data insufficiency, multi-sensor fusion combines time-domain, frequency-domain, and time–frequency features. The framework normalizes the health indicator (HI) using Mahalanobis distance and a sigmoid function, and employs the MD-CUSUM technique for early fault detection. Dynamic threshold updating is achieved through BOX-COX transformation and Chebyshev inequality, establishing confidence intervals. The model demonstrates significant results: lowest RMSE of 5.4267, MAE of 3.7857, highest SOR of 0.90936, HD of 0.93543, PICP of 57.143%, and the narrowest MPIW of 9.45, showing enhanced accuracy and reduced uncertainty. Validation with the XJTU-SY bearing dataset confirms improved performance.

Prediction of Cancer Proneness under Influence of X-rays with Four DNA Mutability and/or Three Cellular Proliferation Assays.

All these assays were applied to skin untransformed fibroblasts derived from eight major cancer syndromes characterized by their excess of relative cancer risk (ERR). Significant correlations with ERR were found between hyper-recombination assessed by the plasmid assay and G2/M arrest and described a third-degree polynomial ERR function and a sigmoidal ERR function, respectively. The product of the hyper-recombination rate and capacity of proliferation described a linear ERR function that permits one to better discriminate each cancer syndrome. Hyper-recombination and cell proliferation were found to obey differential equations that better highlight the intrinsic bases of cancer formation. Further investigations to verify their relevance for cancer proneness induced by exogenous agents are in progress.

Accounting for the mechanical response of the cell membrane during the uptake of random nanoparticles

In order to improve the efficiency of the delivery of cancer treatments to cancer cells, the cellular uptake of nanoparticles (NPs), used as drug delivery systems, is numerically investigated through a mechanical approach. The objective is to optimize the NP's mechanical and geometrical properties to enhance their entry into cancer cells while avoiding benign ones. In previous studies, these properties are modeled as constant during the process of cellular uptake. However, recent observations of the displacement of the membrane's constituents towards the region in the cell membrane where the uptake of the NPs takes place show that the mechanical properties of the membrane vary during this process. Reason for writing The important contribution of adhesion to the wrapping process is already well documented in literature. It is therefore crucial to model this parameter properly as the conclusions made with a constant adhesion model may not be accurate compared to reality. Methodology Based on the existing knowledge on the reaction of membrane constituents to interaction with NPs, a 3-parameter sigmoidal function, accounting for the delay, amplitude, and speed of the reaction, has been used to model the evolution of adhesion. A variance-based sensitivity analysis has then been performed in order to quantify the influence of these parameters on the outputs of the model. Results It was found that the introduction of a variable adhesion tends to alter the predictions of endocytosis of NPs. The contribution of the amplitude and delay is respectively 0.32 and 0.43 times as important as that of the NP's aspect ratio, which is the prominent parameter. The influence of the slope of the transition is the least important parameter and does not appear to contribute to endocytosis. Implications Hence, models of the cellular uptake of NPs should use a variable, instead of constant, adhesion in order a representative as possible of the behavior of the cell membrane. The predictions are different from those obtained using a model with constant adhesion.

Modeling and control strategy of small unmanned helicopter rotation based on deep learning

Unmanned Aerial Vehicle (UAV) can serve as a substitute for workers in some hazardous environments, but the presence of ground effects makes UAVs prone to hardware damage during the recycling process. To study the rotation phenomenon of small UAV landing process and propose control strategy, the study completes the rotation modeling of small unmanned helicopter based on deep learning algorithm and proposes the control strategy. The results show that the CNN with ReLU function has the best performance, and the model converges in the 5th iteration with this function, while the model with Sigmoid function converges in the 36th iteration, and the fitting effect of the rotation model constructed by the study is higher than that of the traditional rotation model. The actual trajectory of the research-constructed rotation model starts to coincide with the expected trajectory at the 5th s of the landing process, while the actual trajectory of the Cheeseman-Bennett model starts to coincide with the expected trajectory only at the 26th s of the landing process. Under the control strategy model proposed in the study, the roll angle and pitch angle of UAV are stabilized at 46s, and the fluctuation of yaw angle is also minimal. The rotation model constructed in the study can completely reflect the rotation process of the small UAV, and the designed control system can help the UAV recover stability faster.

Insight of low flammability limit on sustainable aviation fuel blend and prediction by ANN model

Sustainable aviation fuel (SAF) blend has been confirmed to benefit for greenhouse gases reduction, and thus the property of blend fuel should be understanded the detail to support the utilization in aircraft. Low flammability limit (LFL) is a key property of jet fuel which should be sufficiently flammable to burn in combustor of aeroengine and meanwhile should be non-flammable for safety storage in fuel tank of aircraft. LFL of fuel could be influenced by integrating effects including molecule structure, intramolecular chemical bond energy and binding energy of molecule to molecule. Three types of theoretical models, based on different individual view including LFL of every pure hydrocarbon, stoichiometric concentration, and combustion enthalpy, present unsatisfactory simulation results, which can be deduced without integrating all potential influence factors together. The artificial neural network (ANN) approaches have been involved to bridge the relationship of the complex compositions in jet fuels with LFL. For providing adequate and available composition input, the boundary of fuel compositions has been extracted based on constrains of boiling point, flash point and freeze point coupling with statistic petroleum-based jet fuels. By clustering analysis, 43 critical classes of compositions, extracted as surrogate hydrocarbons based on with similar LFL within 1 % deviation, have been deployed as input matrix. ANN-LFL model, trained by only drop-in fuel with feature of Sigmoid function as an activation function, can distinguish drop-in fuel with non-drop-in fuel. ANN LFL model can predict LFL of drop-in fuel with 0.988 accuracy. The predict output value of non-drop-in fuel could present obvious deviation with traditional jet fuel. The optimization methodologies of ANN-LFL model could be improved the understanding of LFL and extend ANN in SAF utilization.

STUDY OF TEMPERATURE FIELDS NEAR A WELL DURING FILTRATION OF VISCOPLASTIC OIL

The work examines the time variation of fluid temperature during non-isothermal filtration of viscoplastic oil near a well. It is known that a characteristic feature of modern oil production is an increase in the share of hard-to-recover oil reserves, which include viscoplastic oils. According to their rheological characteristics, such oils are non-Newtonian liquids and, therefore, their filtration process is not described by Darcy’s linear law. When filtering fluids that are not described by Darcy's law, it is necessary to use nonlinear filtration laws, which make it possible to more adequately describe the processes occurring in productive formations when hydrocarbons are displaced from them. To describe the nonlinear filtration law, empirical formulas obtained by approximating data from laboratory experiments are usually used.In the works of V.V. Devlikamov and his co-authors provide experimental data on the dependence of viscosity on shear stress for high-viscosity oils. They have been shown to be approximated quite well by a sigmoid function. However, the problem then becomes significantly nonlinear. Subsequently, in order to simplify finding a solution to the problem of filtration of high-viscosity oil, experimental data were obtained by V.V. Devlikamov and his co-authors were approximated by broken lines.In this work, the experimental data are approximated by a sigmoid function, which allows for more accurate calculations. Assuming that the viscosity and density of oil do not depend on temperature, which is typical for natural development modes, the results of calculations of temperature changes over time near the well are presented. It is shown that the temperature of viscoplastic oil in the first hours after starting a well differs significantly from the temperature of Newtonian oil (oil with constant viscosity). Thus, if the temperature of Newtonian oil continuously increases, then the temperature of viscoplastic oil immediately after starting a well increases and coincides with the temperature of oil with constant viscosity, then drops and after some time begins to increase again. The work shows that such an unusual change in the temperature of viscoplastic oil over time is associated with a change in the zone of destruction of the formation structure.

Particle swarm optimization for hybrid mutant slime mold: An efficient algorithm for solving the hyperparameters of adaptive Grey-Markov modified model

Thegrey prediction model is a scientific and effective prediction method for small amounts of incomplete data. In this paper, we proposed a particle swarm optimization for mixed mutant slime mold (namely SCPSO) combined with an extended adaptive Grey-Markov modified model called AOPGM(1,1,λ,µ), which constructs an AOPGM(1,1,λ,µ) based hyperparameters optimization for SCPSO. Firstly, the values of the exponential cumulative generation operator coefficient λ and the background value µ of the adaptive Grey-Markov modified model are refined, and λ and µ are also calculated using SCPSO. Secondly, a competitive SCPSO is formed by introducing a good point set, identifying attack strategy, mutation slime mold algorithm, and Sigmoid function into particle swarm optimization. It was also compared with the improved optimization algorithms on the CEC2020 test set and with classical and newer intelligent algorithms on the CEC2022 test set. The results of numerical experiments show that SCPSO can be considered a competitive and promising global optimization algorithm. Finally, SCPSO is used to optimize the hyperparameters of the AOPGM(1,1,λ,µ) and applied to the prediction of oil production and proven reserves in China. In terms of optimizing the hyperparameters of the AOPGM(1,1,λ,µ), the numerical experimental results of this method were better than the prediction results of 7 intelligent algorithms and the results of 4 prediction models. The validity of the proposed methodology was verified through five evaluation indicators.

Investigation of changes in the cumulative number of magnetic anomalies before and after earthquakes using satellite data

Iran is always at risk of destructive earthquakes due to its location on the Alpine-Himalayan seismic belt. Many significant earthquakes have occurred there so far, which have caused a lot of financial and human losses. The plateau consists of a composite system of collision-oblique transpressive fold-and-thrust mountain belts with active reverse and strike-slip faulting, range-and-basin terrains, active subduction zones, recent volcanic activity, variable crust thicknesses and rigidity, and relatively stable aseismic blocks of different dimensions with low.topographic relief and nearly flat areas., as a result, attention to earthquakes and research in its various fields is necessary and inevitable in Iran. This study investigated Iran’s earthquakes from 2014 to 2021 with a magnitude above 5.5. A total of 30 earthquakes were investigated, of which 21 cases were considered due to solar activities and the simultaneous occurrence of some earthquakes. For this purpose, Swarm Alpha and Swarm Charlie satellite data belonging to the European Space Agency have been used vector magnetic field of all the paths that passed near the place of the earthquake was used between two months before the occurrence and one month after it. First, by performing the first stage of processing for the paths that are close to the earthquake in terms of time, the vivid anomalies caused by the event are identified in the y component of the magnetic field. Then, using these anomalies, a threshold value for the standard deviation is defined for each event so that the anomalies with a higher standard deviation than this threshold value are examined and analyzed. Finally, the occurrence of the studied earthquakes was predicted by fitting the sigmoid function to the cumulative number of anomalies. Because the sigmoid function is not a dynamic function, at the same time, a cubic function has been used to estimate the critical time of the system so that when the sigmoid function cannot fit the data, the cubic function can do this. A first-order curve has also been used to show the proper fit of the data with the sigmoid and cubic functions. And also, C-factor has been used to show the strong compatibility of sigmoid and cubic functions with the cumulative number of anomalies. In the last stage, an analysis has been done under the title of “Confutation analysis” in such a way that the said algorithm is applied to the areas where the earthquake did not occur, and it is expected that the cumulative number of anomalies drawn, will be closer to the linear state than when the earthquake occurred. Using the existing database, 90% of the occurrence of earthquakes studied in this study, pre-indication abnormalities are observed a few hours to a month before the incidence of earthquakes.Also, a Confutation analysis was performed on three cases of earthquakes to show that the obtained results were not random and the change in the cumulative number of magnetic anomalies could be caused by the earthquake.

Morph-SSL: Self-Supervision With Longitudinal Morphing for Forecasting AMD Progression From OCT Volumes.

The lack of reliable biomarkers makes predicting the conversion from intermediate to neovascular age-related macular degeneration (iAMD, nAMD) a challenging task. We develop a Deep Learning (DL) model to predict the future risk of conversion of an eye from iAMD to nAMD from its current OCT scan. Although eye clinics generate vast amounts of longitudinal OCT scans to monitor AMD progression, only a small subset can be manually labeled for supervised DL. To address this issue, we propose Morph-SSL, a novel Self-supervised Learning (SSL) method for longitudinal data. It uses pairs of unlabelled OCT scans from different visits and involves morphing the scan from the previous visit to the next. The Decoder predicts the transformation for morphing and ensures a smooth feature manifold that can generate intermediate scans between visits through linear interpolation. Next, the Morph-SSL trained features are input to a Classifier which is trained in a supervised manner to model the cumulative probability distribution of the time to conversion with a sigmoidal function. Morph-SSL was trained on unlabelled scans of 399 eyes (3570 visits). The Classifier was evaluated with a five-fold cross-validation on 2418 scans from 343 eyes with clinical labels of the conversion date. The Morph-SSL features achieved an AUC of 0.779 in predicting the conversion to nAMD within the next 6 months, outperforming the same network when trained end-to-end from scratch or pre-trained with popular SSL methods. Automated prediction of the future risk of nAMD onset can enable timely treatment and individualized AMD management.

Better approximation of sigmoid function for privacy-preserving neural networks

Abstract With the development of artificial intelligence technology, privacy protection has become a prominent research direction. Homomorphic encryption is an important and promising technology in the construction of privacy-preserving neural networks. However, computing nonlinear activation functions remains a challenging task in the context of homomorphic encryption. Considering Sigmoid is one of the most commonly used activation functions, this paper proposes a polynomial approximation method for the Sigmoid function and applies it to the homomorphic encryption scheme of CKKS. We design polynomial approximation functions using Improved Taylor Expansion Segmentation (ITES), Least Squares method (LS), and Random Noise Addition method, respectively. These methods are implemented through the CKKS scheme in the homomorphic encryption library of SEAL. The experimental results show that the ITES achieves high fitting accuracy, reducing the average absolute error by 11.09% and the mean squared error by 38.06% compared to previous fitting methods. Therefore, experiments have identified methods with better fitting results in the CKKS scheme, which can meet the requirements of using homomorphic encryption in deep learning technologies. Finally, our analysis points out that different technologies have unique advantages and adapt to different scenarios, e.g., ITES has higher accuracy, LS has higher performance in scenarios with limited computing resources.

A Novel EPSO Algorithm Based on Shifted Sigmoid Function Parameters for Maximizing the Energy Yield from Photovoltaic Arrays: An Experimental Investigation

This paper presents a new method for maximizing the energy production of photovoltaic (PV) arrays, utilizing the Enhanced Particle Swarm Optimization (EPSO) algorithm. In contrast with the classical PSO, the proposed EPSO algorithm employs shifted sigmoid functions in order to adapt the acceleration coefficients and the inertia weight instead of using constant coefficients. The EPSO algorithm proves its effectiveness even in challenging either condition, such as abrupt variations in solar irradiance or instances of partial shading conditions (PSCs). The proposed EPSO-based Maximum Power Point Tracking (MPPT) controller has undergone testing and a comparative analysis alongside well-known metaheuristic algorithms, which include the PSO, Bat Algorithm (BAT), Grasshopper Optimization Algorithm (GOA), and Grey Wolf Optimizer (GWO). The results obtained through simulations consistently demonstrate that the EPSO-based MPPT method surpasses other metaheuristic approaches in terms of energy yield and the speed at which it tracks the Maximum Power Point (MPP) under PSCs. In addition, the proposed EPSO algorithm demonstrates robustness, maintaining consistent tracking of the MPP during sudden solar radiation changes. Experimental validation on a real PV system affirms these findings, demonstrating that the EPSO algorithm surpasses its counterparts by achieving a high MPPT efficiency of approximately 99.19% in a fast-tracking time of 2.4 s while maintaining minimal power oscillations of around 2.04 %.