def init_mdp_old(self): #get all network nodes node_count = len(traci.junction.getIDList()) mdp = [] for node in range(0, node_count): no = (self.network.getNode(str(node+1))) list_edges = Node.getOutgoing(no) #for each node find all outgoing edges states = [] for l in list_edges: #add the actions (edges) to the states (nodes): edge-id, q-value, reward (travel time on link) states.append([str(Edge.getID(l)), -50*random.random(), -50*random.random()]) mdp.append(states) return mdp
def init_mdp(self): #get all network nodes junction_list = traci.junction.getIDList() usefull = [i for i in junction_list if not ':' in i] mdp = dict([(x, []) for x in usefull]) for node in usefull: no = (self.network.getNode(node)) list_edges = Node.getOutgoing(no) states = dict([(x.getID(), []) for x in list_edges if not ':' in x.getID()]) #for each node find all outgoing edgesTrue for e in list_edges: #add the actions (edges) to the states (nodes): edge-id, q-value, reward (travel time on link) states[e.getID()] = ([10*random.random(), 10*random.random()]) mdp[node] = states return mdp