frozenlake.py

#frozen lake solver 4*4 
#it is a bruteforce approach with the forbidden states explicitly assigned by me
 
import gym
import random
env = gym.make("FrozenLake-v0")
env.reset()
env.render()
reward=0.00

forbidden=[5,7,11,12]

actions = {
    'Left': 0,
    'Down': 1,
    'Right': 2,
    'Up': 3
}
counter=0
bool=True
while(bool):
    counter=counter+1
    winning_sequence=[random.choice(["Left","Down","Right"]),random.choice(["Left","Down","Right"]),random.choice(["Left","Down"]),random.choice(["Left","Down","Right","Up"])]
    for a in winning_sequence:
        new_state, reward, done, info = env.step(actions[a])
        print()
        env.render()
        print("Reward: {:.2f}".format(reward))
        if new_state in forbidden:
            env.reset()
            break
        if new_state==15:
            bool=False
            break
print("no.of attempts",counter)
print("the winning sequence",winning_sequence)