# optimize_gjo.py

import random

def gjo_optimize(search_space, iterations=10):
    best_score = float('-inf')
    best_params = {}

    for i in range(iterations):
        lr = random.uniform(*search_space['learning_rate'])
        batch_size = random.choice(search_space['batch_size'])
        dropout = random.uniform(*search_space['dropout_rate'])

        # Simulate a scoring function (higher is better)
        score = 0.98 - abs(lr - 0.01) * 5 - abs(dropout - 0.2) * 2

        print(f"Iteration {i+1}: LR={lr:.4f}, Batch={batch_size}, Dropout={dropout:.2f}, Score={score:.4f}")

        if score > best_score:
            best_score = score
            best_params = {
                "learning_rate": lr,
                "batch_size": batch_size,
                "dropout_rate": dropout
            }

    return best_params

if __name__ == '__main__':
    param_space = {
        "learning_rate": (0.001, 0.05),
        "batch_size": [32, 64, 96, 128],
        "dropout_rate": (0.1, 0.5)
    }

    best = gjo_optimize(param_space, iterations=10)
    print("\nBest Hyperparameters Found:")
    print(best)