class AssociativeMemory(object):
'''
classdocs
'''
def __init__(self,size):
'''
Constructor
'''
self.size=size
self.outputPattern=zeros(self.size)
self.connectionWeights=zeros((self.size, self.size))
self.network=Network(self.size)
def learn(self,patterns):
for i in range(0,self.size):
for j in range(0,self.size):
if i==j:
self.connectionWeights[i,j]=0
else:
partial=zeros(len(patterns))
for x in range (0,len(patterns)):
partial[x]=(2*patterns[x][i]-1)*(2*patterns[x][j]-1)
self.connectionWeights[i,j]=sum(partial)
print self.connectionWeights
def recall(self,pattern):
self.network.loadInitialState(pattern)
self.network.loadConnectionMatrix(self.connectionWeights)
self.network.iterate()
def testIterate(self):
network = Network(10)
initialStates = random.randint(2, size=10)
arr = random.random_integers(-2000, 2000, size=(10, 10))
synapsis = (arr + arr.T) / 2.0 - diag(diag(arr))
network.loadInitialState(initialStates)
network.loadConnectionMatrix(synapsis)
network.iterate()
pass
class TSP(object):
'''
classdocs
'''
def __init__(self,size):
'''
Constructor
'''
self.numCities=size
self.networkSize=size*size
self.outputRoute=zeros(self.numCities)
self.distances=zeros((self.numCities, self.numCities))
self.network=Network(self.networkSize)
self.connectionWeights=zeros((self.networkSize,self.networkSize))
def delta(self,a,b):
if a==b:
return 1
else:
return 0
def iterate(self):
initialStates = random.randint(-1, 1, size=self.networkSize )
self.network.loadInitialState(initialStates)
self.network.loadConnectionMatrix(self.connectionWeights)
self.network.iterate()
self.outputRoute=self.network.neuronStates
print "\nRoute:"
print self.outputRoute
def learn(self,distances):
A=1
B=1
C=0.5
for x in range(0,self.numCities):
for y in range(0,self.numCities):
for i in range(0,self.numCities):
for j in range(0,self.numCities):
self.connectionWeights[x+i,y+j]=-A*self.delta(x,y)*(1-self.delta(i,j))-B*self.delta(i,j)*(1-self.delta(x,y))-C*self.delta(x,y)*(self.delta(j,i)+self.delta(j,i-1))
print self.connectionWeights