Intelligent home energy management system (IHEMS) manages various home appliances depending on user preferences in order to save energy cost and assure user satisfaction. To realize cooperation between the distribution system operator (DSO) and end-user, a two-stage scheduling optimization model is developed based on the flexibility quantification, with the user privacy in demand-side management in a specific cluster taken into account. It also aids to reduce the network operation cost, and the distribution system's uncertainty caused by wind power is represented using chance constraints. The user uploads the IHEMS-calculated offered flexibility to DSO via cluster; DSO optimizes the optimal flexibility to minimize the operation cost; then, clusters distribute the flexibility requirement based on the flexibility index; finally, the user updates the optimal schedule. The numerical results using MATLAB are provided to show the effectiveness of the proposed approach.

This paper proposed a reinforcement learning (RL) based energy management system of pelagic islands network microgrids (PINMGs) by ship swapping under the influence of environmental impacts. In addition, the day-ahead standard scheduling by proposing a novel method to maximize the usage of renewable energy (RE) proposes the energy-sharing structure between islands. Energy sharing among islands plays an important role in electrifying remote islands, which need energy due to the unavailability of renewable energy resources to meet local demand. The two-stage cooperative multi-agent deep RL has been presented with the deep Q-learning (DQN) based approach with central RL and island agents (IA) distributed over numerous islands to overcome this challenge. The deep RL-based approaches efficiently learn and optimize their behaviors through several epochs compared with other machine learning or conventional methods due to their in-depth learning capability. Hence, the centralized RL-based problem using dueling DQN was solved to schedule charge battery sharing from resource-rich islands (SI) to load island networks (LIN). In addition, the case study has compared the accuracy among different DQN methods and further scheduling based on the dueling DQN because of its accurate tracking. Due to fluctuating swapping demand and charging patterns, the need for LIN is also stochastic. Hence, the simulation results, including energy scheduling through the ship, are validated by maximizing RE usage through sharing between several islands, and the usefulness of the proposed algorithm is verified through state and action perturbation to verify the robustness.

Precise Photovoltaic (PV) power forecasting tools are needed to integrate PV into the new framework of the energy sector. Also, the related intermittent and random nature needs to be appropriately considered. In this context, this paper examines various machine learning algorithms used for one hour ahead forecasting of PV power production. Specifically, the performances of two Long Short-Term Memory (LSTM) recurrent neural networks, a Gradient Boost Machine (GBM) model and an Extreme Gradient Boosting (XGB) model, are compared. Six years of data are retrieved from the 81 kW PV power plant in the Savona campus of the University of Genoa and are used to train and test the algorithms. The performances of all the algorithms are compared over the original dataset, composed of meteorological variables linked with the PV production, a dataset using seasonal and trend decomposition (STL) of the meteorological variables, and some reduced datasets that mimic the situation in which some of the features are not available at least for some time steps. By comparing the results through different case studies, it can be justified that the XGB outperforms the other algorithms and the STL decomposition is helpful in increasing the performance of the model.

This article presents the deep reinforcement learning (DRL) based smart scheduling in intelligent home energy management system (SSIHEMS) for electric vehicles (EVs) scheduling by utilizing the photovoltaic (PV) on the rooftop for economic dispatch problems. Therefore, optimizing home appliances to minimize consumption cost is challenging because of the randomness of electricity prices and poses a challenge for efficient scheduling. The data-driven model-free DRL-based SSIHEMS is utilized to optimize the decision by managing different home appliances and offering appropriate scheduling EVs to overcome the shortcomings. The decision includes the proper scheduling of battery charging, discharging, and EV to reduce the dependency on the electric grid through a collaborative approach. In addition, the proposed work covers designing a gym-based environment that incorporates the states fed to an agent and receives the reward based on the action taken for scheduling. Hence, the case study is performed to validate the proposed approach. It is verified that the decisions for battery charging, discharging, and EV scheduling are managed well through PV generation with respect to time. Furthermore, to verify the robustness and effectiveness, a comparison of different algorithms such as deep Q-network (DQN), double DQN, and dueling DQN.