def RL_forwarding(data, src, dst):
graph = get_dict(data)
A = graph["A"]
Z = graph["Z"]
weight = graph["weight"]
links = graph["links"]
R = initial_R(A,Z,weight,links)
Q = initial_Q(R)
alpha = 0.9 # learning rate
epsilon = 0.8 #greedy policy
n_episodes = 300
return get_result(R,Q,alpha,epsilon,n_episodes,src,dst)
def find_nodes(csv_file, start_node, end_node):
data = pd.read_csv(csv_file)
graph = get_dict(data)
A = graph["A"]
Z = graph["Z"]
weight = graph["weight"]
A_Z_dict = graph["A_Z_dict"]
start = int(start_node)
end = [int(end_node)]
R = initial_R(A, Z, weight, A_Z_dict)
Q_init = initial_Q(R)
Q = copy.deepcopy(Q_init)
# print("The WEIGHT-TABLE from a NODE A TO NODE B : ")
# print(R)
# print("The Q-TABLE after initialisation : ")
# print(Q)
alpha = 0.7 # learning rate
epsilon = 0.1 #greedy policy
n_episodes = 900
# print("The final Q table : ")
# time0 = time.time()
result = get_result(R, Q, alpha, epsilon, n_episodes, start, end)
# print("time is:",time.time() - time0)
# print(result["ends_find"])
# print(result["cost"])
# print(result["routes_number"])
# print(result["all_routes"])
return R, Q_init, result["all_routes"], result["final_Q"]
import pandas as pd
import time
data = pd.read_csv("graph.csv")
graph = get_dict(data)
A = graph["A"]
Z = graph["Z"]
weight = graph["weight"]
A_Z_dict = graph["A_Z_dict"]
##
start = 1
end = [9]
R = initial_R(A, Z, weight, A_Z_dict)
Q = initial_Q(R)
alpha = 0.7 # learning rate
epsilon = 0.1 #greedy policy
n_episodes = 1000
time0 = time.time()
result = get_result(R, Q, alpha, epsilon, n_episodes, start, end)
print("time is:", time.time() - time0)
print(result["ends_find"])
print(result["cost"])
print(result["routes_number"])