def test_getForests(self):
forest = Forest(self.size)
location1 = (1, 3)
location2 = (2, 3)
location3 = (4, 4)
location4 = (1, 1)
location5 = (1, 2)
location6 = (4, 2)
forest.addTree(location1)
forest.addTree(location2)
forest.addTree(location3)
forest.addTree(location4)
forest.addTree(location5)
forest.addTree(location6)
forests = forest.getForests()
num_forests = 3
self.assertEqual(len(forests), num_forests)
class HotModel:
def __init__(self, grid_size, design_parameter):
self.grid_size = grid_size
self.design_parameter = design_parameter
self.prob_scale = self.getCharacteristicScale()
self.norm_const = self.getNormalizationConstant()
self.forest = Forest(self.grid_size)
def getNormalizationConstant(self):
return ((1 - exp(1 / self.prob_scale)) /
(1 - exp(-self.grid_size / self.prob_scale)))**2
def getCharacteristicScale(self):
return self.grid_size / 10
def probability(self, location):
# characteristic scale for the distribution
l = self.prob_scale
x = location[0] + 1
y = location[1] + 1
return self.norm_const * exp(-x / l) * exp(-y / l)
def isFullyPopulated(self):
return self.getNumTrees() >= self.grid_size**2
def getNumTrees(self):
return self.forest.getNumTrees()
def getNormalizedNumTrees(self):
return self.getNumTrees() / self.grid_size**2
def getForestYield(self):
return self.normalizedAverageYieldFrom(self.forest)
def getForestMatrix(self):
matrix = np.copy(self.forest.getForestMatrix())
return matrix
def getForests(self):
return self.forest.getForests()
def normalizedAverageYieldFrom(self, forest):
return self.averageYieldFrom(forest) / self.grid_size**2
def averageYieldFrom(self, forest):
return forest.getNumTrees() - self.averageLightningDamageFrom(forest)
def averageLightningDamageFrom(self, forest):
forest_matrix = forest.getForestMatrix()
damage = 0
for index, forest_size in np.ndenumerate(forest_matrix):
loc_prob = self.probability(index)
damage += forest_size * loc_prob
return damage
def addTree(self):
location = self.getNextTreeLocation()
self.forest.addTree(location)
def getNextTreeLocation(self):
best_location = None
best_yield = -1
for location in self.getFutureSearchLocations():
simulated_forest = self.forest.copy()
simulated_forest.addTree(location)
avg_yield = self.averageYieldFrom(simulated_forest)
if avg_yield > best_yield:
best_yield = avg_yield
best_location = location
if best_location is None:
raise RuntimeError('Next Tree Location Not Found')
return best_location
def getFutureSearchLocations(self):
possible_locations = self.forest.getUnoccupiedLocations()
shuffle(possible_locations)
max_index = min(self.design_parameter, len(possible_locations))
search_locations = possible_locations[0:max_index]
return search_locations