-
Notifications
You must be signed in to change notification settings - Fork 15
/
test_sao.py
527 lines (469 loc) · 19.3 KB
/
test_sao.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
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
from __future__ import annotations
import random
from random import choice
from negmas.sao.negotiators import AspirationNegotiator
import hypothesis.strategies as st
from hypothesis import example, given, settings
from pytest import mark
import negmas
from negmas import (
PolyAspiration,
PresortingInverseUtilityFunction,
all_negotiator_types,
)
from negmas.common import PreferencesChangeType
from negmas.gb.common import ResponseType
from negmas.outcomes import Issue, make_issue
from negmas.outcomes.outcome_space import make_os
from negmas.preferences import LinearAdditiveUtilityFunction
from negmas.preferences import LinearAdditiveUtilityFunction as LUFun
from negmas.preferences.value_fun import AffineFun, IdentityFun, LinearFun
from negmas.sao import EndImmediately, NoneOfferingPolicy, RejectAlways, SAOMechanism
from negmas.sao.common import SAOResponse, SAOState
from negmas.sao.negotiators.base import SAONegotiator
from negmas.sao.negotiators.modular.boa import make_boa
NEGTYPES = all_negotiator_types()
class SmartAspirationNegotiator(SAONegotiator):
_inv = None # The ufun invertor (finds outcomes in a utility range)
_partner_first = None # The best offer of the partner (assumed best for it)
_min = None # The minimum of my utility function
_max = None # The maximum of my utility function
_best = None # The best outcome for me
def __init__(self, *args, **kwargs):
# initialize the base SAONegoiator (MUST be done)
super().__init__(*args, **kwargs)
# Initialize the aspiration mixin to start at 1.0 and concede slowly
self._asp = PolyAspiration(1.0, "boulware")
def on_preferences_changed(self, changes):
# create an initiaze an invertor for my ufun
changes = [_ for _ in changes if _.type not in (PreferencesChangeType.Scale,)]
if not changes:
return
self._inv = PresortingInverseUtilityFunction(self.ufun) # type: ignore
self._inv.init()
# find worst and best outcomes for me
worest, self._best = self.ufun.extreme_outcomes() # type: ignore
# and the correponding utility values
self._min, self._max = self.ufun(worest), self.ufun(self._best) # type: ignore
# MUST call parent to avoid being called again for no reason
super().on_preferences_changed(changes)
def respond(self, state, source: str | None = None):
offer = state.current_offer
if offer is None:
return ResponseType.REJECT_OFFER
# set the partner's first offer when I receive it
if not self._partner_first:
self._partner_first = offer
# accept if the offer is not worse for me than what I would have offered
return super().respond(state, source)
def propose(self, state):
# calculate my current aspiration level (utility level at which I will offer and accept)
a = (self._max - self._min) * self._asp.utility_at( # type: ignore
state.relative_time
) + self._min
# find some outcomes (all if the outcome space is discrete) above the aspiration level
outcomes = self._inv.some((a - 1e-6, self._max + 1e-6), False) # type: ignore
# If there are no outcomes above the aspiration level, offer my best outcome
if not outcomes:
return self._best
# else if I did not recieve anything from the partner, offer any outcome above the aspiration level
if not self._partner_first:
return choice(outcomes)
# otherwise, offer the outcome most similar to the partner's first offer (above the aspiration level)
nearest, ndist = None, float("inf")
for o in outcomes:
d = sum((a - b) * (a - b) for a, b in zip(o, self._partner_first))
if d < ndist:
nearest, ndist = o, d
return nearest
def try_negotiator(cls, replace_buyer=True, replace_seller=True, n_steps=100):
buyer_cls = cls if replace_buyer else AspirationNegotiator
seller_cls = cls if replace_seller else AspirationNegotiator
# create negotiation agenda (issues)
issues = [
make_issue(name="price", values=10),
make_issue(name="quantity", values=(1, 11)),
make_issue(name="delivery_time", values=10),
]
# create the mechanism
session = SAOMechanism(issues=issues, n_steps=n_steps)
# define ufuns
seller_utility = LUFun(
values={ # type: ignore
"price": IdentityFun(),
"quantity": LinearFun(0.2),
"delivery_time": AffineFun(-1, bias=9),
},
weights={"price": 1.0, "quantity": 1.0, "delivery_time": 10.0},
outcome_space=session.outcome_space,
reserved_value=15.0,
).scale_max(1.0)
buyer_utility = LUFun(
values={ # type: ignore
"price": AffineFun(-1, bias=9.0),
"quantity": LinearFun(0.2),
"delivery_time": IdentityFun(),
},
outcome_space=session.outcome_space,
reserved_value=10.0,
).scale_max(1.0)
session.add(buyer_cls(name="buyer"), ufun=buyer_utility) # type: ignore
session.add(seller_cls(name="seller"), ufun=seller_utility) # type: ignore
session.run()
return session
@given(
opp=st.sampled_from(NEGTYPES),
start=st.booleans(),
rejector=st.sampled_from([EndImmediately, RejectAlways]),
)
@example(
opp=negmas.sao.negotiators.timebased.AdditiveFirstFollowingTBNegotiator,
start=True,
rejector=negmas.sao.components.acceptance.EndImmediately,
)
@settings(deadline=500000)
def test_do_nothing_never_gets_agreements(opp, start, rejector):
agent = make_boa(acceptance=rejector(), offering=NoneOfferingPolicy())
issues: list[Issue] = [
make_issue(10, "price"),
make_issue(10, "quantity"),
make_issue(["red", "green", "blue"], "color"),
]
ufuns = [
LinearAdditiveUtilityFunction.random(issues=issues),
LinearAdditiveUtilityFunction.random(issues=issues),
]
session = SAOMechanism(n_steps=1000, issues=issues)
negs = [opp(), agent] if not start else [agent, opp()]
for n, u in zip(negs, ufuns):
session.add(n, preferences=u)
assert session.run().agreement is None
@mark.parametrize(
["factory", "name", "short_name"], [(make_boa, "BOANegotiator", "BOA")]
)
def test_has_correct_type_name(factory, name, short_name):
x = factory()
assert x.type_name == name
assert x.short_type_name == short_name
@mark.repeat(3)
def test_pend_works():
os = make_os(
[
make_issue(10, "price"),
make_issue(10, "quantity"),
make_issue(["red", "green", "blue"], "color"),
]
)
for _ in range(50):
ufuns = [
LinearAdditiveUtilityFunction.random(outcome_space=os, reserved_value=0.0),
LinearAdditiveUtilityFunction.random(outcome_space=os, reserved_value=0.0),
]
n = 1000
f = 0.01
session = SAOMechanism(
n_steps=None, time_limit=None, pend=f / n, outcome_space=os
)
for i, u in enumerate(ufuns):
neg = AspirationNegotiator()
assert session.add(neg, preferences=u) # type: ignore
assert len(session.negotiators) == (i + 1)
assert abs(session.expected_relative_time - (f / (n + 1))) < 1e-8
assert session.expected_remaining_time is None
assert session.expected_remaining_steps is not None
assert abs(session.expected_remaining_steps - n / f) < 4
assert abs(session.relative_time - (f / (n + 1))) < 1e-8
assert session.remaining_steps is None
assert session.remaining_time is None
assert session.state.step <= 10000 * n
assert not session.state.started
agreement = session.run().agreement
assert session.state.started and session.state.ended
if agreement is not None:
break
else:
raise AssertionError("agreement failed in all runs")
def test_pend_per_second_works():
issues: list[Issue] = [
make_issue(10, "price"),
make_issue(10, "quantity"),
make_issue(["red", "green", "blue"], "color"),
]
ufuns = [
LinearAdditiveUtilityFunction.random(issues=issues, reserved_value=0.0),
LinearAdditiveUtilityFunction.random(issues=issues, reserved_value=0.0),
]
n = 10
session = SAOMechanism(
n_steps=None, time_limit=None, pend_per_second=1 / n, issues=issues
)
for u in ufuns:
session.add(AspirationNegotiator(), preferences=u) # type: ignore
assert session.expected_relative_time < 1e-8
assert (
session.expected_remaining_time is not None
and abs(session.expected_remaining_time - n) < 1e-8
)
assert session.expected_remaining_steps is None
assert session.relative_time < 1e-8
assert session.remaining_steps is None
assert session.remaining_time is None
session.run()
assert session.state.time <= 100 * n
@mark.parametrize("s", [1, 3, 10, 101, 1000])
def test_nsteps_apply_as_round(s):
issues: list[Issue] = [
make_issue(10, "price"),
make_issue(10, "quantity"),
make_issue(["red", "green", "blue"], "color"),
]
ufuns = [
LinearAdditiveUtilityFunction.random(issues=issues, reserved_value=0.0),
LinearAdditiveUtilityFunction.random(issues=issues, reserved_value=0.0),
]
session = SAOMechanism(n_steps=s, issues=issues)
for u in ufuns:
assert session.add(AspirationNegotiator(), preferences=u) # type: ignore
assert session.expected_remaining_steps == s
assert session.remaining_steps == s
assert session.current_step == 0
assert abs(session.relative_time - (1.0 / (s + 1))) < 1e-6
assert session.remaining_time is None
session.step()
assert session.current_step == 1
assert session.expected_remaining_steps == (s - 1)
assert session.remaining_steps == s - 1
assert abs(session.relative_time - (2.0 / (s + 1))) < 1e-6
assert session.remaining_time is None
session.run()
ndone = session.current_step
for nid in session.negotiator_ids:
assert len(session.negotiator_offers(nid)) in (ndone, ndone - 1)
assert session.state.step <= s
@mark.parametrize("s", [1, 3, 10, 101, 1000])
def test_nsteps_apply_as_step(s):
issues: list[Issue] = [
make_issue(10, "price"),
make_issue(10, "quantity"),
make_issue(["red", "green", "blue"], "color"),
]
ufuns = [
LinearAdditiveUtilityFunction.random(issues=issues, reserved_value=0.0),
LinearAdditiveUtilityFunction.random(issues=issues, reserved_value=0.0),
]
session = SAOMechanism(n_steps=s, issues=issues, one_offer_per_step=True)
for u in ufuns:
assert session.add(AspirationNegotiator(), preferences=u) # type: ignore
assert session.expected_remaining_steps == s
assert session.remaining_steps == s
assert session.current_step == 0
assert abs(session.relative_time - (1.0 / (s + 1))) < 1e-6
assert session.remaining_time is None
session.step()
assert session.current_step == 1
assert session.expected_remaining_steps == (s - 1)
assert session.remaining_steps == s - 1
assert abs(session.relative_time - (2.0 / (s + 1))) < 1e-6
assert session.remaining_time is None
session.run()
ndone = session.current_step
for nid in session.negotiator_ids:
assert len(session.negotiator_offers(nid)) in (
int(ndone / 2),
int(ndone / 2) + 1,
int((ndone - 1) / 2),
int((ndone - 1) / 2) + 1,
)
assert session.state.step <= s
def test_basic_sao():
n_steps = 100
issues: list[Issue] = [
make_issue(10, "price"),
make_issue(5, "quantity"),
make_issue(["red", "green", "blue"], "color"),
]
os = make_os(issues)
ufuns = [
LinearAdditiveUtilityFunction.random(outcome_space=os, reserved_value=0.0),
LinearAdditiveUtilityFunction.random(outcome_space=os, reserved_value=0.0),
LinearAdditiveUtilityFunction.random(outcome_space=os, reserved_value=0.0),
]
session = SAOMechanism(n_steps=n_steps, outcome_space=os, one_offer_per_step=True)
agents = [AspirationNegotiator() for _ in range(len(ufuns))]
for u, a in zip(ufuns, agents):
assert session.add(a, ufun=u) # type: ignore
# offers = [os.random_outcome() for _ in range(n_steps)]
assert session.expected_remaining_steps == n_steps
assert session.remaining_steps == n_steps
assert session.current_step == 0
assert abs(session.relative_time - (1.0 / (n_steps + 1))) < 1e-6
assert session.remaining_time is None
assert not session.state.started and not session.state.running
for i in range(n_steps):
if not session.step().running:
break
assert (
session.state.started and session.state.running
), f"{session.state=}\n{session.extended_trace=}"
assert (
session.current_step == i + 1
), f"{session.state=}\n{session.extended_trace=}"
assert session.expected_remaining_steps == (
n_steps - i - 1
), f"{session.state=}\n{session.extended_trace=}"
assert (
session.remaining_steps == n_steps - i - 1
), f"{session.state=}\n{session.extended_trace=}"
assert (
abs(session.relative_time - ((i + 2) / (n_steps + 1))) < 1e-6
), f"{session.state=}\n{session.extended_trace=}"
assert session.remaining_time is None
assert session.state.started and not session.state.running
assert session.state.step <= n_steps
def test_basic_sao_with_action():
n_steps = 50
issues: list[Issue] = [
make_issue(10, "price"),
make_issue(5, "quantity"),
make_issue(["red", "green", "blue"], "color"),
]
os = make_os(issues)
ufuns = [
LinearAdditiveUtilityFunction.random(outcome_space=os, reserved_value=0.0),
LinearAdditiveUtilityFunction.random(outcome_space=os, reserved_value=0.0),
LinearAdditiveUtilityFunction.random(outcome_space=os, reserved_value=0.0),
]
session = SAOMechanism(n_steps=n_steps, outcome_space=os, one_offer_per_step=True)
agents = [AspirationNegotiator() for _ in range(len(ufuns))]
ids = [_.id for _ in agents]
for u, a in zip(ufuns, agents):
assert session.add(a, ufun=u) # type: ignore
offers = [os.random_outcome() for _ in range(n_steps)]
assert session.expected_remaining_steps == n_steps
assert session.remaining_steps == n_steps
assert session.current_step == 0
assert abs(session.relative_time - (1.0 / (n_steps + 1))) < 1e-6
assert session.remaining_time is None
assert not session.state.started and not session.state.running
for i in range(n_steps):
action = None
pass_action = random.random() < 0.5
if pass_action:
ids = session.next_negotitor_ids()
assert len(ids) == 1
action = {ids[0]: SAOResponse(ResponseType.REJECT_OFFER, offers[i])}
if not session.step(action).running:
break
if pass_action:
state: SAOState = session.state # type: ignore
assert state.current_offer == offers[i]
assert state.current_proposer == ids[0]
assert (
session.state.started and session.state.running
), f"{session.state=}\n{session.extended_trace=}"
assert (
session.current_step == i + 1
), f"{session.state=}\n{session.extended_trace=}"
assert session.expected_remaining_steps == (
n_steps - i - 1
), f"{session.state=}\n{session.extended_trace=}"
assert (
session.remaining_steps == n_steps - i - 1
), f"{session.state=}\n{session.extended_trace=}"
assert (
abs(session.relative_time - ((i + 2) / (n_steps + 1))) < 1e-6
), f"{session.state=}\n{session.extended_trace=}"
assert session.remaining_time is None
assert session.state.started and (
not session.state.running or session.state.step >= n_steps
), f"Did not finish running:\n{session.extended_trace}"
assert (
session.state.step <= n_steps
), f"Ran for too long {session.state.step} but max expected is {n_steps} steps:\n{session.extended_trace}"
class MyNeg(AspirationNegotiator):
def __init__(self, *args, **kwargs):
super().__init__(*args, **kwargs)
self.negstarted = False
self.negended = False
def on_negotiation_start(self, *args, **kwargs):
assert not self.negstarted
self.negstarted = True
super().on_negotiation_start(*args, **kwargs)
def respond(self, *args, **kwargs):
return ResponseType.REJECT_OFFER
def on_negotiation_end(self, *args, **kwargs):
assert not self.negended
self.negended = True
super().on_negotiation_end(*args, **kwargs)
def test_hidden_time_works_and_no_call_repetitions():
time, hidden = 18000, 30
issues: list[Issue] = [
make_issue(10, "price"),
make_issue(5, "quantity"),
make_issue(["red", "green", "blue"], "color"),
]
os = make_os(issues)
ufuns = [
LinearAdditiveUtilityFunction.random(outcome_space=os, reserved_value=0.0),
LinearAdditiveUtilityFunction.random(outcome_space=os, reserved_value=0.0),
]
session = SAOMechanism(
time_limit=time,
n_steps=None,
hidden_time_limit=hidden,
outcome_space=os,
one_offer_per_step=False,
ignore_negotiator_exceptions=False,
)
agents = [MyNeg() for _ in range(len(ufuns))]
for u, a in zip(ufuns, agents):
assert session.add(a, ufun=u) # type: ignore
state = session.run()
assert state.timedout
assert 0.85 * hidden <= state.time <= hidden * 1.3
def test_smart_asipration():
try_negotiator(SmartAspirationNegotiator)
class RTRecorder(SAONegotiator):
def __init__(self, *args, **kwargs):
self.records = []
super().__init__(*args, **kwargs)
def __call__(self, state: SAOState) -> SAOResponse:
self.records.append(
(
state.step,
state.relative_time,
state.time,
# ((state.step + 1) / (self.nmi.n_steps + 1) if state.step > 0 else 0.0)
(state.step + 1) / (self.nmi.n_steps + 1) if self.nmi.n_steps else -1,
)
)
return SAOResponse(ResponseType.REJECT_OFFER, self.nmi.random_outcome())
def test_relative_time():
time, hidden = float("inf"), float("inf")
issues: list[Issue] = [
make_issue(10, "price"),
make_issue(5, "quantity"),
make_issue(["red", "green", "blue"], "color"),
]
os = make_os(issues)
ufuns = [
LinearAdditiveUtilityFunction.random(outcome_space=os, reserved_value=0.0),
LinearAdditiveUtilityFunction.random(outcome_space=os, reserved_value=0.0),
]
session = SAOMechanism(
time_limit=time,
n_steps=10,
hidden_time_limit=hidden,
outcome_space=os,
one_offer_per_step=False,
ignore_negotiator_exceptions=False,
)
agents = [RTRecorder() for _ in range(len(ufuns))]
for u, a in zip(ufuns, agents):
assert session.add(a, ufun=u) # type: ignore
session.run()
for agent in agents:
for step, relative_time, time, expected_rt in agent.records:
assert (
abs(relative_time - expected_rt) < 1e-5
), f"{(step, relative_time, time, expected_rt)}"