Control Strategy for Continuous Mountain Car

Problem

The Continuous Mountain Car problem involves controlling a car to reach a target on a hill.

• Given: Car position and velocity on an uneven terrain

• Objective: Reach the target position at the top of the hill

• Constraints: Force limited to [-1.0, 1.0]

Algorithm Design Task

Design a control strategy that selects appropriate force based on car state.

• Inputs: Position, velocity, last action

• Outputs: Force to apply (float in [-1.0, 1.0])

Evaluation

• Environment: Gym MountainCarContinuous-v0

• Fitness: Episode outcome (succeeded/failed)

Template

template_program = '''
import numpy as np

def choose_action(pos: float, v: float, last_action: float) -> float:
    """
    Args:
        pos: Car's position, between [-1.2, 0.6]
        v: Car's velocity, between [-0.07, 0.07]
        last_action: Car's last force, between [-1, 1]

    Return:
         A float representing the force to be applied.
         The value should be in the range of [-1.0, 1.0].
    """
    return np.random.uniform(-1.0, 1.0)
'''

task_description = ("Implement a function that designs a novel strategy function that "
                    "guides the car along an uneven road towards a target.")

Configuration Parameters

Parameter	Description	Default
max_steps	Maximum steps per episode	500
timeout_seconds	Maximum evaluation time	20

Example Usage

from llm4ad.task.machine_learning.car_mountain_continue import CarMountainCEvaluation

evaluator = CarMountainCEvaluation()

def choose_action(pos, v, last_action):
    if pos < 0:
        return -1.0  # Go left to build momentum
    else:
        return 1.0  # Go right towards goal

result = evaluator.evaluate_program('', choose_action)