OPTIMIZING MACHINE LEARNING-DRIVEN MOBILE CHARGING STATIONS IN POWER-CONSTRAINED ENVIRONMENTS

Abstract

Mobile charging stations serve as essential sources of power for a range of devices, such as electric cars, medical equipment, and communication devices in order to maintain the functionality and security of vital activities. A distinct set of optimization problems, such as resource allocation, forecasting model development, power management, cost optimization, and so on, arise when mobile charging stations are deployed and managed in energy-constrained environments. This work addresses the optimization difficulties in the deployment and management of mobile charging stations by utilizing state-of-the-art strategies to improve the operational efficiency and resilience. The system emphasized on employing supervised learning as a kind of machine learning(ML), in which the model was tested and trained on a labeled dataset derived from a range of energy sources. These sources include solar energy systems, generators, and portable battery packs to get around these optimization challenges. The support vector machine (SVM) and multilayer neural networks (MLNP) along with Naïve Bayesian (NB) optimization strategies were developed to boost the performance of mobile charging stations. Artificial neural network (ANN) outperformed the SVM by a significant margin when selecting the target variable. The black-box feature incorporated into the framework reduced error probability and promoted high standards, all while increasing the efficiency and reliability of the model learning process. This study provides an approach for places with limited energy resources and promotes the shift to more environmentally friendly means of transportation while simultaneously enhancing the accessibility, efficiency, and dependability of electrical vehicle charging services.