11. Testing in GAMA GUI

By Killian Trouillet

---

From Headless to GUI

During training, we used GAMA in headless mode for speed. Now that the model is trained, we switch to the GAMA GUI to watch the agent in action.

Key Difference

Mode	Port	Speed	Display	Use
Headless	1001	Fast (no rendering)	None	Training
GUI	1000	Slow (renders display)	Yes	Testing / Visualization

---

Starting GAMA GUI

1. Open GAMA normally (double-click the application)
2. The GUI server runs on port 1000 by default — no extra setup needed

---

The Test Script

Loading the Trained Model

from stable_baselines3 import PPO

model = PPO.load("saved_models/ppo_forager")

This loads the neural network weights saved during training.

Connecting to GUI

The only change from training: the port number.

env = None
for attempt in range(5):
    try:
        env = gym.make(
            "gama_gymnasium_env/GamaEnv-v0",
            gaml_experiment_path="path/to/forager_gym.gaml",
            gaml_experiment_name="gym_env",
            gama_ip_address="localhost",
            gama_port=1000,
        )
        break
    except Exception as e:
        print(f"GAMA GUI not ready. Retrying in 5s...")
        time.sleep(5)

if env is None:
    raise RuntimeError("Failed to connect to GAMA GUI after several attempts.")

Deterministic Evaluation

action, _ = model.predict(obs, deterministic=True)

During training, the policy adds noise to explore new strategies. During testing, we set deterministic=True to use the pure learned policy — no randomness, just the best action the network has learned.

Slow Motion

time.sleep(0.1)  # 100ms delay between steps

We add a small delay between steps so the movement is visible in the GAMA display.

---

Complete Test Script

Note: The GAML model is unchanged from Step 9. See the Complete Model section there for the full forager_gym.gaml.

"""
Smart Forager - Test Script (GUI Visualization)
Author: Killian Trouillet

Usage:
    1. Open GAMA GUI (default port 1000)
    2. Run this script: python test_forager.py
"""

import asyncio
from pathlib import Path
import gymnasium as gym
import gama_gymnasium
from stable_baselines3 import PPO
import numpy as np
import time


GAML_FILE = str(Path(__file__).parent / "forager_gym.gaml")
EXPERIMENT_NAME = "gym_env"
GAMA_PORT = 1000
MODEL_PATH = str(Path(__file__).parent / "saved_models" / "ppo_forager")
NUM_EPISODES = 5
STEP_DELAY = 0.1


async def test():
    print("=" * 50)
    print("  Smart Forager - PPO Test (GUI)")
    print("=" * 50)
    
    print(f"Loading model from: {MODEL_PATH}")
    model = PPO.load(MODEL_PATH)
    print("Model loaded successfully!")
    
    # ---- Create the environment (connected to GAMA GUI) ----
    # We use a retry loop because GAMA GUI takes time to initialize the simulation
    env = None
    for attempt in range(5):
        try:
            print(f"Connecting to GAMA GUI (attempt {attempt + 1}/5)...")
            env = gym.make(
                "gama_gymnasium_env/GamaEnv-v0",
                gaml_experiment_path=GAML_FILE,
                gaml_experiment_name=EXPERIMENT_NAME,
                gama_ip_address="localhost",
                gama_port=GAMA_PORT,
            )
            break
        except Exception as e:
            if "Unable to find" in str(e) or "NOTREADY" in str(e):
                print(f"GAMA simulation not ready yet. Waiting 5s...")
                time.sleep(5)
            else:
                raise e
    
    if env is None:
        print("Failed to connect to GAMA GUI after several attempts.")
        return
    
    print(f"\nRunning {NUM_EPISODES} test episodes...\n")
    
    all_rewards = []
    all_steps = []
    
    for episode in range(NUM_EPISODES):
        obs, info = env.reset()
        time.sleep(1)
        
        total_reward = 0.0
        step = 0
        done = False
        
        while not done:
            action, _ = model.predict(obs, deterministic=True)
            obs, reward, terminated, truncated, info = env.step(action)
            total_reward += reward
            step += 1
            done = terminated or truncated
            time.sleep(STEP_DELAY)
        
        all_rewards.append(total_reward)
        all_steps.append(step)
        
        status = "FOUND FOOD!" if terminated else "TIMEOUT"
        print(f"  Episode {episode + 1}/{NUM_EPISODES}: "
              f"{status} | Steps: {step} | Reward: {total_reward:.1f}")
    
    env.close()
    
    print(f"\n{'=' * 50}")
    print(f"  Test Results Summary")
    print(f"{'=' * 50}")
    print(f"  Episodes:      {NUM_EPISODES}")
    print(f"  Avg Reward:    {np.mean(all_rewards):.1f}")
    print(f"  Avg Steps:     {np.mean(all_steps):.0f}")
    print(f"  Success Rate:  "
          f"{sum(1 for r in all_rewards if r > 50) / NUM_EPISODES * 100:.0f}%")
    print(f"{'=' * 50}")


if __name__ == "__main__":
    asyncio.run(test())

Running the Test

cd models/gym
python test_forager.py

Watch the GAMA display — the blue forager should navigate smoothly around obstacles to reach the green food!

---

Expected Console Output

==================================================
  Smart Forager - PPO Test (GUI)
==================================================
Loading model from: saved_models/ppo_forager
Model loaded successfully!

Running 5 test episodes...

  Episode 1/5: FOUND FOOD! | Steps: 47 | Reward: 93.2
  Episode 2/5: FOUND FOOD! | Steps: 52 | Reward: 91.8
  Episode 3/5: FOUND FOOD! | Steps: 45 | Reward: 94.1
  Episode 4/5: FOUND FOOD! | Steps: 49 | Reward: 92.5
  Episode 5/5: FOUND FOOD! | Steps: 51 | Reward: 91.4

==================================================
  Test Results Summary
==================================================
  Episodes:      5
  Avg Reward:    92.6
  Avg Steps:     49
  Success Rate:  100%
==================================================

---

Summary

Concept	Implementation
Continuous world	No grid, free {x, y} movement in 100×100 space
GAMA↔Python bridge	GymAgent species + gama-gymnasium WebSocket
Neural network policy	PPO("MlpPolicy", ...) via Stable Baselines3
Headless training	gama-headless.bat -socket 1001 → fast, no GUI
GUI testing	PPO.load() + connect to GAMA GUI (port 1000)
Action space	Box([-1,-1], [1,1]) — continuous velocity (dx, dy)
Observation space	Box([0]×13, [1]×13) — position + food direction + 8 sensors
Reward shaping	Distance delta bonus + food reward + living penalty

Key GAML Concepts Used

global, species, reflex, action, aspect, experiment, parameter, geometry, rectangle, circle, line, distance_to, towards, intersects, inter, cos, sin, list, map, loop, draw

Key Python Concepts Used

gymnasium, gym.make(), env.reset(), env.step(), PPO, MlpPolicy, model.learn(), model.save(), model.load(), model.predict(), asyncio, BaseCallback

Key Files

File	Description
models/gym/forager_gym.gaml	GAMA model with GymAgent bridge
models/gym/train_forager.py	PPO training script (headless)
models/gym/test_forager.py	Testing script (GUI visualization)