-
Notifications
You must be signed in to change notification settings - Fork 0
/
example.py
401 lines (364 loc) · 14.9 KB
/
example.py
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
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
# -*- coding: utf-8 -*-
"""Markov Decision Process (MDP) Toolbox: ``example`` module
=========================================================
The ``example`` module provides functions to generate valid MDP transition and
reward matrices.
Available functions
-------------------
forest
A simple forest management example
rand
A random example
"""
# Copyright (c) 2011-2014 Steven A. W. Cordwell
# Copyright (c) 2009 INRA
#
# All rights reserved.
#
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions are met:
#
# * Redistributions of source code must retain the above copyright notice,
# this list of conditions and the following disclaimer.
# * Redistributions in binary form must reproduce the above copyright notice,
# this list of conditions and the following disclaimer in the documentation
# and/or other materials provided with the distribution.
# * Neither the name of the <ORGANIZATION> nor the names of its contributors
# may be used to endorse or promote products derived from this software
# without specific prior written permission.
#
# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
# AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
# IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
# ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE
# LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
# CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
# SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
# INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
# CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
# ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
# POSSIBILITY OF SUCH DAMAGE.
import numpy as _np
import scipy.sparse as _sp
def forest(S=3, r1=4, r2=2, p=0.1, is_sparse=False):
"""Generate a MDP example based on a simple forest management scenario.
This function is used to generate a transition probability
(``A`` × ``S`` × ``S``) array ``P`` and a reward (``S`` × ``A``) matrix
``R`` that model the following problem. A forest is managed by two actions:
'Wait' and 'Cut'. An action is decided each year with first the objective
to maintain an old forest for wildlife and second to make money selling cut
wood. Each year there is a probability ``p`` that a fire burns the forest.
Here is how the problem is modelled.
Let {0, 1 . . . ``S``-1 } be the states of the forest, with ``S``-1 being
the oldest. Let 'Wait' be action 0 and 'Cut' be action 1.
After a fire, the forest is in the youngest state, that is state 0.
The transition matrix ``P`` of the problem can then be defined as follows::
| p 1-p 0.......0 |
| . 0 1-p 0....0 |
P[0,:,:] = | . . 0 . |
| . . . |
| . . 1-p |
| p 0 0....0 1-p |
| 1 0..........0 |
| . . . |
P[1,:,:] = | . . . |
| . . . |
| . . . |
| 1 0..........0 |
The reward matrix R is defined as follows::
| 0 |
| . |
R[:,0] = | . |
| . |
| 0 |
| r1 |
| 0 |
| 1 |
R[:,1] = | . |
| . |
| 1 |
| r2 |
Parameters
---------
S : int, optional
The number of states, which should be an integer greater than 1.
Default: 3.
r1 : float, optional
The reward when the forest is in its oldest state and action 'Wait' is
performed. Default: 4.
r2 : float, optional
The reward when the forest is in its oldest state and action 'Cut' is
performed. Default: 2.
p : float, optional
The probability of wild fire occurence, in the range ]0, 1[. Default:
0.1.
is_sparse : bool, optional
If True, then the probability transition matrices will be returned in
sparse format, otherwise they will be in dense format. Default: False.
Returns
-------
out : tuple
``out[0]`` contains the transition probability matrix P and ``out[1]``
contains the reward matrix R. If ``is_sparse=False`` then P is a numpy
array with a shape of ``(A, S, S)`` and R is a numpy array with a shape
of ``(S, A)``. If ``is_sparse=True`` then P is a tuple of length ``A``
where each ``P[a]`` is a scipy sparse CSR format matrix of shape
``(S, S)``; R remains the same as in the case of ``is_sparse=False``.
Examples
--------
>>> import mdptoolbox.example
>>> P, R = mdptoolbox.example.forest()
>>> P
array([[[ 0.1, 0.9, 0. ],
[ 0.1, 0. , 0.9],
[ 0.1, 0. , 0.9]],
<BLANKLINE>
[[ 1. , 0. , 0. ],
[ 1. , 0. , 0. ],
[ 1. , 0. , 0. ]]])
>>> R
array([[ 0., 0.],
[ 0., 1.],
[ 4., 2.]])
>>> Psp, Rsp = mdptoolbox.example.forest(is_sparse=True)
>>> len(Psp)
2
>>> Psp[0]
<3x3 sparse matrix of type '<... 'numpy.float64'>'
with 6 stored elements in Compressed Sparse Row format>
>>> Psp[1]
<3x3 sparse matrix of type '<... 'numpy.int64'>'
with 3 stored elements in Compressed Sparse Row format>
>>> Rsp
array([[ 0., 0.],
[ 0., 1.],
[ 4., 2.]])
>>> (Psp[0].todense() == P[0]).all()
True
>>> (Rsp == R).all()
True
"""
assert S > 1, "The number of states S must be greater than 1."
assert (r1 > 0) and (r2 > 0), "The rewards must be non-negative."
assert 0 <= p <= 1, "The probability p must be in [0; 1]."
# Definition of Transition matrix
if is_sparse:
P = []
rows = list(range(S)) * 2
cols = [0] * S + list(range(1, S)) + [S - 1]
vals = [p] * S + [1-p] * S
P.append(_sp.coo_matrix((vals, (rows, cols)), shape=(S, S)).tocsr())
rows = list(range(S))
cols = [0] * S
vals = [1] * S
P.append(_sp.coo_matrix((vals, (rows, cols)), shape=(S, S)).tocsr())
else:
P = _np.zeros((2, S, S))
P[0, :, :] = (1 - p) * _np.diag(_np.ones(S - 1), 1)
P[0, :, 0] = p
P[0, S - 1, S - 1] = (1 - p)
P[1, :, :] = _np.zeros((S, S))
P[1, :, 0] = 1
# Definition of Reward matrix
R = _np.zeros((S, 2))
R[S - 1, 0] = r1
R[:, 1] = _np.ones(S)
R[0, 1] = 0
R[S - 1, 1] = r2
return(P, R)
def _randDense(states, actions, mask):
"""Generate random dense ``P`` and ``R``. See ``rand`` for details.
"""
# definition of transition matrix : square stochastic matrix
P = _np.zeros((actions, states, states))
# definition of reward matrix (values between -1 and +1)
R = _np.zeros((actions, states, states))
temp = _np.random.randint(1,2)
#print(temp)
for action in range(actions):
for state in range(states):
# create our own random mask if there is no user supplied one
if mask is None:
m = _np.random.random(states)
r = _np.random.random()
m[m <= r] = 0
m[m > r] = 1
elif mask.shape == (actions, states, states):
m = mask[action][state] # mask[action, state, :]
else:
m = mask[state]
# Make sure that there is atleast one transition in each state
if m.sum() == 0:
m[_np.random.randint(0, states)] = 1
P[action][state] = _np.random.random(states)
P[action][state][state] = temp
P[action][state] = P[action][state] / P[action][state].sum()
R[action][state] = (2 * _np.random.random(states) -
_np.ones(states, dtype=int))
return(P, R)
def _randSparse(states, actions, mask):
"""Generate random sparse ``P`` and ``R``. See ``rand`` for details.
"""
# definition of transition matrix : square stochastic matrix
P = [None] * actions
# definition of reward matrix (values between -1 and +1)
R = [None] * actions
for action in range(actions):
# it may be more efficient to implement this by constructing lists
# of rows, columns and values then creating a coo_matrix, but this
# works for now
PP = _sp.dok_matrix((states, states))
RR = _sp.dok_matrix((states, states))
for state in range(states):
if mask is None:
m = _np.random.random(states)
m[m <= 2/3.0] = 0
m[m > 2/3.0] = 1
elif mask.shape == (actions, states, states):
m = mask[action][state] # mask[action, state, :]
else:
m = mask[state]
n = int(m.sum()) # m[state, :]
if n == 0:
m[_np.random.randint(0, states)] = 1
n = 1
# find the columns of the vector that have non-zero elements
nz = m.nonzero()
if len(nz) == 1:
cols = nz[0]
else:
cols = nz[1]
vals = _np.random.random(n)
vals = vals / vals.sum()
reward = 2*_np.random.random(n) - _np.ones(n)
PP[state, cols] = vals
RR[state, cols] = reward
# PP.tocsr() takes the same amount of time as PP.tocoo().tocsr()
# so constructing PP and RR as coo_matrix in the first place is
# probably "better"
P[action] = PP.tocsr()
R[action] = RR.tocsr()
return(P, R)
def rand(S, A, is_sparse=False, mask=None):
"""Generate a random Markov Decision Process.
Parameters
----------
S : int
Number of states (> 1)
A : int
Number of actions (> 1)
is_sparse : bool, optional
False to have matrices in dense format, True to have sparse matrices.
Default: False.
mask : array, optional
Array with 0 and 1 (0 indicates a place for a zero probability), shape
can be ``(S, S)`` or ``(A, S, S)``. Default: random.
Returns
-------
out : tuple
``out[0]`` contains the transition probability matrix P and ``out[1]``
contains the reward matrix R. If ``is_sparse=False`` then P is a numpy
array with a shape of ``(A, S, S)`` and R is a numpy array with a shape
of ``(S, A)``. If ``is_sparse=True`` then P and R are tuples of length
``A``, where each ``P[a]`` is a scipy sparse CSR format matrix of shape
``(S, S)`` and each ``R[a]`` is a scipy sparse csr format matrix of
shape ``(S, 1)``.
Examples
--------
>>> import numpy, mdptoolbox.example
>>> numpy.random.seed(0) # Needed to get the output below
>>> P, R = mdptoolbox.example.rand(4, 3)
>>> P
array([[[ 0.21977283, 0.14889403, 0.30343592, 0.32789723],
[ 1. , 0. , 0. , 0. ],
[ 0. , 0.43718772, 0.54480359, 0.01800869],
[ 0.39766289, 0.39997167, 0.12547318, 0.07689227]],
<BLANKLINE>
[[ 1. , 0. , 0. , 0. ],
[ 0.32261337, 0.15483812, 0.32271303, 0.19983549],
[ 0.33816885, 0.2766999 , 0.12960299, 0.25552826],
[ 0.41299411, 0. , 0.58369957, 0.00330633]],
<BLANKLINE>
[[ 0.32343037, 0.15178596, 0.28733094, 0.23745272],
[ 0.36348538, 0.24483321, 0.16114188, 0.23053953],
[ 1. , 0. , 0. , 0. ],
[ 0. , 0. , 1. , 0. ]]])
>>> R
array([[[-0.23311696, 0.58345008, 0.05778984, 0.13608912],
[-0.07704128, 0. , -0. , 0. ],
[ 0. , 0.22419145, 0.23386799, 0.88749616],
[-0.3691433 , -0.27257846, 0.14039354, -0.12279697]],
<BLANKLINE>
[[-0.77924972, 0. , -0. , -0. ],
[ 0.47852716, -0.92162442, -0.43438607, -0.75960688],
[-0.81211898, 0.15189299, 0.8585924 , -0.3628621 ],
[ 0.35563307, -0. , 0.47038804, 0.92437709]],
<BLANKLINE>
[[-0.4051261 , 0.62759564, -0.20698852, 0.76220639],
[-0.9616136 , -0.39685037, 0.32034707, -0.41984479],
[-0.13716313, 0. , -0. , -0. ],
[ 0. , -0. , 0.55810204, 0. ]]])
>>> numpy.random.seed(0) # Needed to get the output below
>>> Psp, Rsp = mdptoolbox.example.rand(100, 5, is_sparse=True)
>>> len(Psp), len(Rsp)
(5, 5)
>>> Psp[0]
<100x100 sparse matrix of type '<... 'numpy.float64'>'
with 3296 stored elements in Compressed Sparse Row format>
>>> Rsp[0]
<100x100 sparse matrix of type '<... 'numpy.float64'>'
with 3296 stored elements in Compressed Sparse Row format>
>>> # The number of non-zero elements (nnz) in P and R are equal
>>> Psp[1].nnz == Rsp[1].nnz
True
"""
# making sure the states and actions are more than one
assert S > 1, "The number of states S must be greater than 1."
assert A > 1, "The number of actions A must be greater than 1."
# if the user hasn't specified a mask, then we will make a random one now
if mask is not None:
# the mask needs to be SxS or AxSxS
try:
assert mask.shape in ((S, S), (A, S, S)), "'mask' must have " \
"dimensions S×S or A×S×S."
except AttributeError:
raise TypeError("'mask' must be a numpy array or matrix.")
# generate the transition and reward matrices based on S, A and mask
if is_sparse:
P, R = _randSparse(S, A, mask)
else:
P, R = _randDense(S, A, mask)
return(P, R)
def small():
"""A very small Markov decision process.
The probability transition matrices are::
| | 0.5 0.5 | |
| | 0.8 0.2 | |
P = | |
| | 0.0 1.0 | |
| | 0.1 0.9 | |
The reward matrix is::
R = | 5 10 |
| -1 2 |
Returns
=======
out : tuple
``out[0]`` is a numpy array of the probability transition matriices.
``out[1]`` is a numpy arrray of the reward matrix.
Examples
========
>>> import mdptoolbox.example
>>> P, R = mdptoolbox.example.small()
>>> P
array([[[ 0.5, 0.5],
[ 0.8, 0.2]],
<BLANKLINE>
[[ 0. , 1. ],
[ 0.1, 0.9]]])
>>> R
array([[ 5, 10],
[-1, 2]])
"""
P = _np.array([[[0.5, 0.5], [0.8, 0.2]], [[0, 1], [0.1, 0.9]]])
R = _np.array([[5, 10], [-1, 2]])
return(P, R)