def bestFitNodesCO(m1, m2, children, gr1, gr2, n):
if m1 and m2:
gr1.create(m1, n)
gr2.create(m2, n)
f1 = fitness.computeFitness(gr1)
f2 = fitness.computeFitness(gr2)
nodes1 = gr1.getNodes()
nodes2 = gr2.getNodes()
child1 = []
child2 = []
if f1 > f2:
for nd in nodes1[:3 * len(nodes1)/4]:
child1.extend(nd.getEdgeList())
for nd in nodes2[3 * len(nodes2)/4:]:
child1.extend(nd.getEdgeList())
for nd in nodes1:
child2.extend(nd.getEdgeList())
else:
for nd in nodes2[:3 * len(nodes2)/4]:
child1.extend(nd.getEdgeList())
for nd in nodes1[3 * len(nodes1)/4:]:
child1.extend(nd.getEdgeList())
for nd in nodes2:
child2.extend(nd.getEdgeList())
children.append(child1)
children.append(child2)
return
def bestFitEdgesCO(m1, m2, children, gr1, gr2, n):
if m1 and m2:
gr1.create(m1, n)
gr2.create(m2, n)
f1 = fitness.computeFitness(gr1)
f2 = fitness.computeFitness(gr2)
edges1 = gr1.getEdgeList()
edges2 = gr2.getEdgeList()
child1 = []
child2 = []
if f1 > f2:
child1.extend(edges1[:3 * len(edges1)/4])
child1.extend(edges2[3 * len(edges2)/4:])
child2.extend(edges1)
else:
child1.extend(edges2[:3 * len(edges2)/4])
child1.extend(edges1[3 * len(edges1)/4:])
child2.extend(edges2)
children.append(child1)
children.append(child2)
return
def rouletteBasedSelection(num, population, matingpool, params, op):
params = params.split()
## d = False
## if params[0] == '1':
## d = True
## n = params[1]
## w = False
## if params[6] == 'w':
## w = True
n = int(params[0])
cumulativefs = []
fs = []
#gr = graph.graph.graph(directed = d, weighted = w)
gr = graph.graph.graph()
el = population[0]
gr.create(el, n)
f = fitness.computeFitness(gr)
cumulativefs.append(f)
fs.append(f)
gr.reset()
ms = 0
for el in population[1:]:
gr.create(el, n)
f = fitness.computeFitness(gr)
cumulativefs.append(cumulativefs[-1] + f)
fs.append(f)
gr.reset()
av = cumulativefs[-1] / len(cumulativefs)
op.write("Average Fitness: " + str(av) + '\n')
op.write("Max fitness: " + str(max(fs)) + '\n')
op.write('fittest: ' + str(population[fs.index(max(fs))]) + '\n')
op.write("Min fitness: " + str(min(fs)) + '\n\n')
for i in range(num):
rnd = random.uniform(0, cumulativefs[-1])
index = bisect.bisect_left(cumulativefs, rnd)
matingpool.append(population[index])
del fs[:]
del cumulativefs[:]
return
