Offline Q-Learning Controller for Hybrid PV–Battery System in MATLAB/Simulinkk

Design and implement an offline tabular Q-learning (Q-table) controller in MATLAB/Simulink for a hybrid PV–battery system feeding a buck converter. The controller should accurately regulate the converter’s output voltage under varying irradiance, load, and battery SOC. The project will be completed in two phases: 1. Develop a complete Simulink model of the hybrid PV–battery–buck system, generate an offline dataset of transitions, train the Q-table, and implement a MATLAB Function block for closed-loop voltage regulation using the greedy Q-table controller. 2. Enhance the controller with physics-informed reinforcement learning by incorporating buck converter equations into the reward function and state features. Demonstrate voltage regulation under varying conditions and provide brief documentation covering encoding, reward design, and usage. Deliverables: - Complete Simulink model of the hybrid PV–battery–buck system. - Offline dataset of transitions and trained Q-table. - MATLAB Function block implementing the greedy Q-table controller. - Physics-informed reward/state features based on buck equations. - Demonstration of voltage regulation under varying conditions. - Brief documentation of encoding, reward design, and usage.