alex/copycat
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Merge branch 'feature-temperature-effect-analysis' into revision-2.0

Browse Source
This commit is contained in:
LSaldyt
2017-12-04 12:09:51 -07:00
parent 3d9079cfbd 98c2b497a6
commit 47af738f7c
2 changed files with 40 additions and 14 deletions
Download Patch File Download Diff File Expand all files Collapse all files

17
copycat/problem.py
View File

@ -5,6 +5,12 @@ from pprint import pprint
class Problem: class Problem:
def __init__(self, initial, modified, target, iterations, distributions=None, formulas=None): def __init__(self, initial, modified, target, iterations, distributions=None, formulas=None):
self.formulas = formulas self.formulas = formulas
if formulas is not None:
assert hasattr(Copycat(), 'temperature')
else:
if hasattr(Copycat(), 'temperature'):
self.formulas = set(Copycat().temperature.adj_formulas())
print(self.formulas)
self.initial = initial self.initial = initial
self.modified = modified self.modified = modified
self.target = target self.target = target
@ -14,8 +20,7 @@ class Problem:
self.distributions = self.solve() self.distributions = self.solve()
else: else:
self.distributions = distributions self.distributions = distributions
if formulas is not None: print(self.formulas)
assert hasattr(Copycat().workspace, 'temperature')
def test(self, comparison, expected=None): def test(self, comparison, expected=None):
print('-' * 120) print('-' * 120)
@ -41,8 +46,8 @@ class Problem:
copycat = Copycat() copycat = Copycat()
answers = dict() answers = dict()
if self.formulas == None: if self.formulas == None:
if hasattr(copycat.workspace, 'temperature'): if hasattr(copycat, 'temperature'):
formula = copycat.workspace.temperature.getAdj() formula = copycat.temperature.getAdj()
else: else:
formula = None formula = None
answers[formula] = copycat.run(self.initial, answers[formula] = copycat.run(self.initial,
@ -50,12 +55,14 @@ class Problem:
self.target, self.target,
self.iterations) self.iterations)
else: else:
print(self.formulas)
for formula in self.formulas: for formula in self.formulas:
copycat.temperature.useAdj(formula) copycat.temperature.useAdj(formula)
answers[formulas] = copycat.run(self.initial, answers[formula] = copycat.run(self.initial,
self.modified, self.modified,
self.target, self.target,
self.iterations) self.iterations)
print('Done with {}'.format(formula))
return answers return answers
def generate(self): def generate(self):

37
copycat/temperature.py
View File

@ -24,19 +24,19 @@ def _entropy(temp, prob):
f = (c + 1) * prob f = (c + 1) * prob
return -f * math.log2(f) return -f * math.log2(f)
def _weighted(temp, prob, s, u, alpha=1, beta=1): def _weighted(temp, s, u):
weighted = (temp / 100) * s + ((100 - temp) / 100) * u weighted = (temp / 100) * s + ((100 - temp) / 100) * u
return weighted return weighted
def _weighted_inverse(temp, prob): def _weighted_inverse(temp, prob):
iprob = 1 - prob iprob = 1 - prob
return _weighted(temp, prob, iprob, prob) return _weighted(temp, iprob, prob)
def _fifty_converge(temp, prob): # Uses .5 instead of 1-prob def _fifty_converge(temp, prob): # Uses .5 instead of 1-prob
return _weighted(temp, prob, .5, prob) return _weighted(temp, .5, prob)
def _soft_curve(temp, prob): # Curves to the average of the (1-p) and .5 def _soft_curve(temp, prob): # Curves to the average of the (1-p) and .5
return min(1, _weighted(temp, prob, (1.5-prob)/2, prob)) return min(1, _weighted(temp, (1.5-prob)/2, prob))
def _weighted_soft_curve(temp, prob): # Curves to the weighted average of the (1-p) and .5 def _weighted_soft_curve(temp, prob): # Curves to the weighted average of the (1-p) and .5
weight = 100 weight = 100
@ -49,25 +49,25 @@ def _weighted_soft_curve(temp, prob): # Curves to the weighted average of the (1
def _alt_fifty(temp, prob): def _alt_fifty(temp, prob):
s = .5 s = .5
u = prob ** 2 if prob < .5 else math.sqrt(prob) u = prob ** 2 if prob < .5 else math.sqrt(prob)
return _weighted(temp, prob, s, u) return _weighted(temp, s, u)
def _averaged_alt(temp, prob): def _averaged_alt(temp, prob):
s = (1.5 - prob)/2 s = (1.5 - prob)/2
u = prob ** 2 if prob < .5 else math.sqrt(prob) u = prob ** 2 if prob < .5 else math.sqrt(prob)
return _weighted(temp, prob, s, u) return _weighted(temp, s, u)
def _working_best(temp, prob): def _working_best(temp, prob):
s = .5 # convergence s = .5 # convergence
r = 1.05 # power r = 1.05 # power
u = prob ** r if prob < .5 else prob ** (1/r) u = prob ** r if prob < .5 else prob ** (1/r)
return _weighted(temp, prob, s, u) return _weighted(temp, s, u)
def _soft_best(temp, prob): def _soft_best(temp, prob):
s = .5 # convergence s = .5 # convergence
r = 1.05 # power r = 1.05 # power
u = prob ** r if prob < .5 else prob ** (1/r) u = prob ** r if prob < .5 else prob ** (1/r)
return _weighted(temp, prob, s, u) return _weighted(temp, s, u)
def _parameterized_best(temp, prob): def _parameterized_best(temp, prob):
# (D$66/100)*($E$64*$B68 + $G$64*$F$64)/($E$64 + $G$64)+((100-D$66)/100)*IF($B68 > 0.5, $B68^(1/$H$64), $B68^$H$64) # (D$66/100)*($E$64*$B68 + $G$64*$F$64)/($E$64 + $G$64)+((100-D$66)/100)*IF($B68 > 0.5, $B68^(1/$H$64), $B68^$H$64)
@ -78,7 +78,24 @@ def _parameterized_best(temp, prob):
s = (alpha * prob + beta * s) / (alpha + beta) s = (alpha * prob + beta * s) / (alpha + beta)
r = 1.05 r = 1.05
u = prob ** r if prob < .5 else prob ** (1/r) u = prob ** r if prob < .5 else prob ** (1/r)
return _weighted(temp, prob, s, u) return _weighted(temp, s, u)
def _meta(temp, prob):
r = _weighted(temp, 1, 2) # Make r a function of temperature
s = .5
u = prob ** r if prob < .5 else prob ** (1/r)
return _weighted(temp, s, u)
def _meta_parameterized(temp, prob):
r = _weighted(temp, 1, 2) # Make r a function of temperature
alpha = 5
beta = 1
s = .5
s = (alpha * prob + beta * s) / (alpha + beta)
u = prob ** r if prob < .5 else prob ** (1/r)
return _weighted(temp, s, u)
def _none(temp, prob): def _none(temp, prob):
return prob return prob
@ -99,6 +116,8 @@ class Temperature(object):
'best' : _working_best, 'best' : _working_best,
'sbest' : _soft_best, 'sbest' : _soft_best,
'pbest' : _parameterized_best, 'pbest' : _parameterized_best,
'meta' : _meta,
'pmeta' : _meta_parameterized,
'none' : _none} 'none' : _none}
self.diffs = 0 self.diffs = 0
self.ndiffs = 0 self.ndiffs = 0