Deep Convolutional Neural Networks (DCNNs) have shown remarkable success in image classification tasks, but optimizing their hyperparameters can be challenging due to their complex structure. This paper develops the Adaptive Habitat Biogeography-Based Optimizer (AHBBO) for tuning the hyperparameters of DCNNs in image classification tasks. In complicated optimization problems, the BBO suffers from premature convergence and insufficient exploration. In this regard, an adaptable habitat is presented as a solution to these problems; it would permit variable habitat sizes and regulated mutation. Better optimization performance and a greater chance of finding high-quality solutions across a wide range of problem domains are the results of this modification's increased exploration and population diversity. AHBBO is tested on 53 benchmark optimization functions and demonstrates its effectiveness in improving initial stochastic solutions and converging faster to the optimum. Furthermore, DCNN-AHBBO is compared to 23 well-known image classifiers on nine challenging image classification problems and shows superior performance in reducing the error rate by up to 5.14%. Our proposed algorithm outperforms 13 benchmark classifiers in 87 out of 95 evaluations, providing a high-performance and reliable solution for optimizing DNNs in image classification tasks. This research contributes to the field of deep learning by proposing a new optimization algorithm that can improve the efficiency of deep neural networks in image classification.
This study investigated the chemical properties of peat microparticles modified asphalt (Pt.M.A.). The originality of the study resides in the examination of the chemical characteristics of peat microparticles (Pt.) modified asphalt (Pt. M.A.) utilising FTIR, SEM, SFE, and XRD methodologies. This encompasses Fourier transform infrared spectroscopy (FTIR), scanning electron microscopy (SEM), surface free energy (SFE), and X-ray diffraction (XRD). Initially, FTIR examined the functional groups of both unaltered and altered asphalt binders. The SEM images reveal improved compatibility, showcasing superior diffusion of the modifier across the asphalt. A further critical factor is that improved adhesion properties, according to the SFE study, indicate that modified binders generally offer more SFE compared to unmodified binders. The XRD measurements revealed a semi-crystalline structure in the Pt. modifier and an amorphous structure in the basal asphalt binder. The integration of Pt. into the asphalt cement resulted in modifications to the phases of both constituents, culminating in the emergence of a new semi-crystalline phase inside the modified asphalt binder. These data suggest that peat microparticles (Pt.) can improve the efficacy of asphalt binders by enhancing compatibility, adhesion, and resistance to ageing.
The maximum power delivered by a photovoltaic system is greatly influenced by atmospheric conditions such as irradiation and temperature and by surrounding objects like trees, raindrops, tall buildings, animal droppings, and clouds. The partial shading caused by these surrounding objects and the rapidly changing atmospheric parameters make maximum power point tracking (MPPT) challenging. This paper proposes a hybrid MPPT algorithm that combines the benefits of the salp swarm algorithm (SSA) and hill climbing (HC) techniques. As long as the rate of change of irradiance does not exceed a specific limit, the HC mode is applied to track the global maximum power point (GMPP). Once a high rate of change in irradiation is detected, the SSA mode is activated. Moreover, the proposed algorithm employs the concept of boundary conditions to handle fast and slow fluctuating irradiance patterns. A comprehensive comparative evaluation of the proposed hybrid SSA-HC with state-of-the-art MPPT algorithms has been undertaken. Four distinct cases have been examined, including irradiance conditions with varying rates of change and partial shading conditions. The proposed hybrid SSA-HC algorithm has been validated and tested using a developed hardware setup, simulated in MATLAB for solar photovoltaic (PV) systems, and compared with standard SSA and HC. The performance of the tracking capability of this proposed hybrid technique at both steady-state and dynamic conditions under rapid and gradual irradiance changes demonstrates its superiority over recent state-of-the-art algorithms.