def runAlgorithms(self, algorithms, real_mode=False):
self.algorithms = algorithms
for alg_name, alg in list(algorithms.items()):
self.AlgReward[alg_name] = []
self.averageReward[alg_name] = []
self.resultRecord()
for iter_ in range(self.iterations):
for alg_name, algorithm in list(algorithms.items()):
S = algorithm.decide()
# print("----------the ", iter_ + 1, " round---------")
# print('choose set:', S)
if real_mode:
reward, live_edges, live_nodes = runReal_DC(self.G, S)
else:
if alg_name[:2] == 'DI':
reward, live_edges, live_nodes = runDC_DILinUCB(self.G, self.Prob, S)
else:
reward, live_edges, live_nodes = runDC(self.G, self.Prob, S)
algorithm.updateParameters(S, live_nodes, live_edges, iter_)
print("rewards: " + str(reward))
self.AlgReward[alg_name].append(reward)
# self.averageReward[alg_name].append("%.2f" % np.mean(self.AlgReward[alg_name][0:iter_ + 1]))
self.resultRecord(iter_)
def simulate(self, S):
reward, live_edges, live_nodes = runDC(self.G, self.real_P, S)
return live_edges, reward
def simulate(self, S):
reward, live_edges, live_nodes = runDC(self.G, self.real_P, S)
# print("----------------rewards :", reward)
return live_edges, reward
def simulate(self, S):
reward, live_edges, live_nodes = runDC(self.G, self.real_P, S)
# observed_probabilities : (node n, index i): reward
# print("----------------simulation rewards :", reward)
return live_edges, reward