main.cpp

#include <iostream>
#include <iomanip>
#include "RL_Lib.h"

using namespace RLLib;

int main() {
    std::cout << "========================================" << std::endl;
    std::cout << "RL Library v" << RL_LIB_VERSION << std::endl;
    std::cout << "========================================" << std::endl << std::endl;

    // 1. Initialize Environment
    std::cout << "[1] Initializing GridWorld Environment..." << std::endl;
    GridWorld env(5, 5, 4, 4);
    std::cout << "    Grid Size: " << env.getWidth() << "x" << env.getHeight() << std::endl;
    std::cout << "    Total States: " << env.getNumStates() << std::endl;
    std::cout << "    Goal Position: (" << env.getGoalPosition().first << ", " 
              << env.getGoalPosition().second << ")" << std::endl << std::endl;

    // 2. Initialize Agent (Q-Learner)
    std::cout << "[2] Initializing Q-Learning Agent..." << std::endl;
    QLearner agent(0.1, 0.99, 0.1);  // alpha=0.1, gamma=0.99, epsilon=0.1
    std::cout << "    Learning Rate (alpha): " << agent.alpha << std::endl;
    std::cout << "    Discount Factor (gamma): " << agent.gamma << std::endl;
    std::cout << "    Exploration Rate (epsilon): " << agent.epsilon << std::endl << std::endl;

    // 3. Training Loop
    std::cout << "[3] Starting Training..." << std::endl;
    int num_episodes = 100;
    int max_steps = 50;

    for (int episode = 0; episode < num_episodes; ++episode) {
        env.reset();
        int state = env.getState();
        double episode_reward = 0.0;

        for (int step = 0; step < max_steps; ++step) {
            // Choose action using epsilon-greedy strategy
            int action = agent.chooseAction(state, GridWorld::NUM_ACTIONS);
            
            // Execute action and get reward
            double reward = env.step(action);
            int next_state = env.getState();
            episode_reward += reward;

            // Update Q-values
            agent.update(state, action, reward, next_state, GridWorld::NUM_ACTIONS);

            state = next_state;

            // Check if terminal state
            if (env.isTerminal()) {
                break;
            }
        }

        // Print progress every 10 episodes
        if ((episode + 1) % 10 == 0) {
            std::cout << "    Episode " << std::setw(3) << (episode + 1) 
                      << " | Reward: " << std::fixed << std::setprecision(2) 
                      << episode_reward << std::endl;
        }
    }

    std::cout << "\n[4] Training Complete!" << std::endl;
    std::cout << "========================================" << std::endl;

    return 0;
}