rl-tp2-dynamic-programming/exercices.py at main · guhur/rl-tp2-dynamic-programming · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
"""
Ce fichier contient des exercices à compléter sur la programmation dynamique.
Il est évalué automatiquement avec pytest, vous pouvez le lancer avec la
commande `pytest exercices.py`.
"""
import os
import random
import typing as t

import gym
import numpy as np
import pytest
from gym import spaces

from dynamic_programming import MDP, GridWorldEnv, StochasticGridWorldEnv
from dynamic_programming.domino_paving import domino_paving
from dynamic_programming.fibonacci import fibonacci, fibonacci_memo
from dynamic_programming.values_iteration import (
    grid_world_value_iteration, mdp_value_iteration,
    stochastic_grid_world_value_iteration)


# Tests pour l'exercice 1
def test_mdp():
    mdp = MDP()
    assert mdp.P[0][0] == (1, -1, False)
    assert mdp.P[0][1] == (0, -1, False)
    assert mdp.P[1][0] == (0, -1, False)
    assert mdp.P[1][1] == (2, -1, False)
    assert mdp.P[2][0] == (2, 0, False)
    assert mdp.P[2][1] == (0, -1, False)

    mdp.reset()
    ret = mdp.step(0)
    assert ret[0] in [0, 1, 2]
    assert ret[1] in [0, -1]
    assert ret[2] in [True, False]
    assert isinstance(ret[3], dict)


def test_mdp_value_iteration(max_iter: int = 1000):
    mdp = MDP()
    values = mdp_value_iteration(mdp, max_iter=max_iter, gamma=1.0)
    assert np.allclose(values, [-2, -1, 0]), print(values)
    values = mdp_value_iteration(mdp, max_iter=max_iter, gamma=0.9)
    assert np.allclose(values, [-1.9, -1, 0])


def test_grid_world_value_iteration(max_iter=1000):
    env = GridWorldEnv()

    values = grid_world_value_iteration(env, max_iter, gamma=1.0)
    solution = np.array(
        [
            [1.0, 1.0, 1.0, 0.0],
            [1.0, 0.0, 1.0, 0.0],
            [1.0, 1.0, 1.0, 1.0],
            [1.0, 1.0, 1.0, 1.0],
        ]
    )

    assert np.allclose(values, solution), print(values)

    values = grid_world_value_iteration(env, max_iter, gamma=0.9)
    solution = np.array(
        [
            [0.81, 0.9, 1.0, 0.0],
            [0.729, 0.0, 0.9, 0.0],
            [0.6561, 0.729, 0.81, 0.729],
            [0.59049, 0.6561, 0.729, 0.6561],
        ]
    )
    assert np.allclose(values, solution)


def test_stochastic_grid_world_value_iteration(max_iter=1000):
    env = StochasticGridWorldEnv()

    values = stochastic_grid_world_value_iteration(env, max_iter=max_iter, gamma=1.0)
    solution = np.array(
        [
            [1.0, 1.0, 1.0, 0.0],
            [1.0, 0.0, 1.0, 0.0],
            [1.0, 1.0, 1.0, 1.0],
            [1.0, 1.0, 1.0, 1.0],
        ]
    )
    assert np.allclose(values, solution), print("  ", values)

    values = stochastic_grid_world_value_iteration(env, max_iter=max_iter, gamma=0.9)
    solution = np.array(
        [
            [0.77495822, 0.87063224, 0.98343293, 0.0],
            [0.68591168, 0.0, 0.77736888, 0.0],
            [0.60732544, 0.60891859, 0.68418232, 0.60570595],
            [0.54079452, 0.54500607, 0.60570595, 0.53914484],
        ]
    )
    assert np.allclose(values, solution), print("  ", values)


# Tests pour l'exercice 1
@pytest.mark.parametrize(
    "n,expected",
    [
        (0, 0),
        (1, 1),
        (2, 1),
        (3, 2),
        (5, 5),
        (10, 55),
        (20, 6765),
    ],
)
def test_fibonacci(n, expected):
    assert fibonacci(n) == expected


# Tests pour l'exercice 2
@pytest.mark.parametrize(
    "n,expected",
    [
        (0, 0),
        (1, 1),
        (2, 1),
        (3, 2),
        (5, 5),
        (10, 55),
        (20, 6765),
    ],
)
def test_fibonacci_memo(n, expected):
    assert fibonacci_memo(n) == expected


# Tests pour l'exercice 3
@pytest.mark.parametrize(
    "n,expected",
    [
        (1, 0),
        (2, 3),
        (3, 0),
        (4, 11),
        (5, 0),
        (6, 41),
        (7, 0),
        (8, 153),
        (9, 0),
        (10, 571),
    ],
)
def test_domino_paving(n, expected):
    assert domino_paving(n) == expected


def test_wall():
    env = GridWorldEnv()
    for i in range(2):
        env.step(0)
    old_position = env.current_position
    env.step(3)
    assert old_position == env.current_position, print(
        env.current_position, old_position
    )