class CachingEnv:
def __init__(self):
self.rsu_num = RSU_NUM
self.request_num = REQUEST_NUM
self.region_num = REGION_NUM
with open('experimentData/RegionRsu/8region_4rsu.txt', 'r') as f:
line = f.readline()
strlist = str.split(line)
intlist = list(map(int, strlist))
self.region_rsu = np.array(intlist)
with open('experimentData/RsuConnect/4rsuConnect.txt', 'r') as f:
temparr = []
for line in f:
strlist = str.split(line)
intlist = list(map(int, strlist))
temparr.append(intlist)
self.rsu_connect = np.array(temparr)
with open('experimentData/RequestSize/10request.txt', 'r') as f:
line = f.readline()
strlist = str.split(line)
intlist = list(map(int, strlist))
self.request_size = np.array(intlist)
self.requestGenerator = RequestGenerator(
5, 60, 8, [35, 17, 11, 9, 6, 6, 5, 4, 4, 3])
#self.requestFile = open('experimentData/RegionRequest/8region_10request.txt')
self.cache_state = np.zeros((RSU_NUM, REQUEST_NUM), dtype=int)
self.region_request = self.requestGenerator.generateRegionRequestMatrix(
)
self.rsu_capcity = np.ones(RSU_NUM) * DEFAULT_RSU_CAPCITY
self.rsu_residual_capcity = np.ones(RSU_NUM) * DEFAULT_RSU_CAPCITY
self.request_popularity = np.zeros(REQUEST_NUM, dtype=int)
self.n_actions = 4
self.n_features = RSU_NUM + REQUEST_NUM
self.index_of_core = 2
self.time1 = 1
self.time2 = 2
self.time3 = 3
self.time4 = 4
def reset(self):
self.cache_state = np.zeros((RSU_NUM, REQUEST_NUM), dtype=int)
self.rsu_residual_capcity = np.ones(RSU_NUM) * DEFAULT_RSU_CAPCITY
self.request_popularity = np.zeros(REQUEST_NUM, dtype=int)
#self.requestFile = open('experimentData/RegionRequest/8region_10request.txt')
# temparr = []
# for i in range(self.region_num):
# line = self.requestFile.readline()
# strlist = str.split(line)
# intlist = list(map(int, strlist))
# temparr.append(intlist)
#self.region_request = np.array(temparr)
self.region_request = self.requestGenerator.generateRegionRequestMatrix(
)
return np.concatenate(
[self.rsu_residual_capcity, self.request_popularity])
def step(self, action):
# temparr = []
# for i in range(self.region_num):
# line = self.requestFile.readline()
# strlist = str.split(line)
# intlist = list(map(int, strlist))
# temparr.append(intlist)
# self.region_request = np.array(temparr)
currentRequestCount, hitCacheCount = self.hitCacheCount()
self.calculatePopularity()
if action == 0: #global popular content in core rsu
popularId = np.argmax(self.request_popularity)
if self.cache_state[self.index_of_core][
popularId] == 0 and self.rsu_residual_capcity[
self.index_of_core] >= self.request_size[popularId]:
self.cache_state[self.index_of_core][popularId] = 1
self.rsu_residual_capcity[
self.index_of_core] -= self.request_size[popularId]
elif action == 1: #current popular content in local rsu
currentContentPopularity = {}
for i in range(REQUEST_NUM):
currentContentPopularity[i] = 0
for i in range(REGION_NUM):
for j in range(REQUEST_NUM):
if self.region_request[i][j] == 1:
currentContentPopularity[j] += 1
popularId = max(currentContentPopularity,
key=currentContentPopularity.get)
tempRequest = self.region_request[:, popularId]
for i in range(REGION_NUM):
if tempRequest[i] == 1:
tempRSUId = self.region_rsu[i]
if self.cache_state[tempRSUId][
popularId] == 0 and self.rsu_residual_capcity[
tempRSUId] >= self.request_size[popularId]:
self.cache_state[tempRSUId][popularId] = 1
self.rsu_residual_capcity[
tempRSUId] -= self.request_size[popularId]
elif action == 2: #current popular content in core rsu
currentContentPopularity = {}
for i in range(REQUEST_NUM):
currentContentPopularity[i] = 0
for i in range(REGION_NUM):
for j in range(REQUEST_NUM):
if self.region_request[i][j] == 1:
currentContentPopularity[j] += 1
popularId = max(currentContentPopularity,
key=currentContentPopularity.get)
if self.cache_state[self.index_of_core][
popularId] == 0 and self.rsu_residual_capcity[
self.index_of_core] >= self.request_size[popularId]:
self.cache_state[self.index_of_core][popularId] = 1
self.rsu_residual_capcity[
self.index_of_core] -= self.request_size[popularId]
else: #random cache
tempRSUId = random.randint(0, RSU_NUM - 1)
tempRequestId = random.randint(0, REQUEST_NUM - 1)
if self.cache_state[tempRSUId][
tempRequestId] == 0 and self.rsu_residual_capcity[
tempRSUId] >= self.request_size[tempRequestId]:
self.cache_state[tempRSUId][tempRequestId] = 1
self.rsu_residual_capcity[tempRSUId] -= self.request_size[
tempRequestId]
observation_ = np.concatenate(
[self.rsu_residual_capcity, self.request_popularity])
store_cost = (np.sum(self.rsu_capcity) - np.sum(
self.rsu_residual_capcity)) / np.sum(self.rsu_capcity)
tranTime_sum = 0
for i in range(REGION_NUM):
for j in range(REQUEST_NUM):
tempRSUId = self.region_rsu[i]
if self.cache_state[tempRSUId][j] == 1:
tranTime_sum += self.time2
else:
tranTime_sum += self.time4
trans_cost = tranTime_sum / (self.time4 * REQUEST_NUM * REGION_NUM)
cost = store_cost + trans_cost
reward = -cost
self.region_request = self.requestGenerator.generateRegionRequestMatrix(
)
return observation_, reward, currentRequestCount, hitCacheCount
def hitCacheCount(self):
requestCount = 0
cacheHitCount = 0
for i in range(len(self.region_request)):
for j in range(len(self.region_request[0])):
if self.region_request[i][j] == 1:
requestCount += 1
correspondRsuId = self.region_rsu[i]
if self.cache_state[correspondRsuId][j] == 1:
cacheHitCount += 1
return requestCount, cacheHitCount
def calculatePopularity(self):
for i in range(REGION_NUM):
for j in range(REQUEST_NUM):
if self.region_request[i][j] == 1:
self.request_popularity[j] += 1
class CachingEnv:
def __init__(self):
self.rsu_num = RSU_NUM
self.request_num = REQUEST_NUM
self.region_num = REGION_NUM
self.car_num = CAR_NUM
with open('experimentData/RegionRsu/8region_4rsu.txt', 'r') as f:
line = f.readline()
strlist = str.split(line)
intlist = list(map(int, strlist))
self.region_rsu = np.array(intlist)
with open('experimentData/RsuConnect/4rsuConnect.txt', 'r') as f:
temparr = []
for line in f:
strlist = str.split(line)
intlist = list(map(int, strlist))
temparr.append(intlist)
self.rsu_connect = np.array(temparr)
with open('experimentData/RequestSize/25request.txt', 'r') as f:
line = f.readline()
strlist = str.split(line)
intlist = list(map(int, strlist))
self.request_size = np.array(intlist)
with open('experimentData/RegionNeighbor/8regions.txt', 'r') as f:
temparr = []
for line in f:
strlist = str.split(line)
intlist = list(map(int, strlist))
temparr.append(intlist)
self.region_neighbor = np.array(temparr)
#self.requestGenerator = RequestGenerator(5, 60, 8, [52, 20, 12, 9, 7])
#self.requestGenerator = RequestGenerator(5, 60, 8,[35,17,11,9,6,6,5,4,4,3])
#self.requestGenerator = RequestGenerator(5, 60, 8, [31,15,10,8,6,5,4,4,3,3,3,3,2,2,1])
#self.requestGenerator = RequestGenerator(5, 60, 8, [28,14,10,7,6,5,4,4,3,3,2,2,2,2,2,2,1,1,1,1])
#self.requestGenerator = RequestGenerator(5, 60, 8,[27, 13, 9, 7, 5, 4, 4, 3, 3, 3, 2, 2, 2, 2, 2, 2, 2, 1, 1, 1, 1, 1, 1,1, 1])
self.requestGenerator = RequestGenerator(5, 60, 8, [
1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
1, 1, 1
])
#self.requestFile = open('experimentData/RegionRequest/8region_10request.txt')
self.cache_state = np.zeros((RSU_NUM, REQUEST_NUM), dtype=int)
self.car_cache_state = np.zeros((CAR_NUM, REQUEST_NUM), dtype=int)
self.region_request = self.requestGenerator.generateRegionRequestMatrix(
)
self.rsu_capcity = np.ones(RSU_NUM) * DEFAULT_RSU_CAPCITY
self.car_capcity = np.ones(CAR_NUM) * DEFAULT_CAR_CAPCITY
self.rsu_residual_capcity = np.ones(RSU_NUM) * DEFAULT_RSU_CAPCITY
self.car_residual_capcity = np.ones(CAR_NUM) * DEFAULT_CAR_CAPCITY
self.request_popularity = np.zeros(REQUEST_NUM, dtype=int)
self.car_location = []
for i in range(CAR_NUM):
temp = random.randint(0, REGION_NUM - 1)
self.car_location.append(temp)
self.n_actions = 5
self.n_features = RSU_NUM + CAR_NUM + REQUEST_NUM
self.index_of_core = 2
self.time1 = 1
self.time2 = 2
self.time3 = 3
self.time4 = 4
def reset(self):
self.cache_state = np.zeros((RSU_NUM, REQUEST_NUM), dtype=int)
self.car_cache_state = np.zeros((CAR_NUM, REQUEST_NUM), dtype=int)
self.rsu_residual_capcity = np.ones(RSU_NUM) * DEFAULT_RSU_CAPCITY
self.car_residual_capcity = np.ones(CAR_NUM) * DEFAULT_CAR_CAPCITY
self.request_popularity = np.zeros(REQUEST_NUM, dtype=int)
#self.requestFile = open('experimentData/RegionRequest/8region_10request.txt')
# temparr = []
# for i in range(self.region_num):
# line = self.requestFile.readline()
# strlist = str.split(line)
# intlist = list(map(int, strlist))
# temparr.append(intlist)
#self.region_request = np.array(temparr)
self.region_request = self.requestGenerator.generateRegionRequestMatrix(
)
self.car_location = []
for i in range(CAR_NUM):
temp = random.randint(0, REGION_NUM - 1)
self.car_location.append(temp)
return np.concatenate([
self.rsu_residual_capcity, self.car_residual_capcity,
self.request_popularity
])
def step(self, action):
# temparr = []
# for i in range(self.region_num):
# line = self.requestFile.readline()
# strlist = str.split(line)
# intlist = list(map(int, strlist))
# temparr.append(intlist)
# self.region_request = np.array(temparr)
currentRequestCount, hitCacheCount = self.hitCacheCount()
self.calculatePopularity()
if action == 0:
currentContentPopularity = {}
for i in range(REQUEST_NUM):
currentContentPopularity[i] = 0
for i in range(REGION_NUM):
for j in range(REQUEST_NUM):
if self.region_request[i][j] == 1:
currentContentPopularity[j] += 1
popularId = max(currentContentPopularity,
key=currentContentPopularity.get)
#popularId = np.argmax(self.request_popularity)
for tempRSUId in range(RSU_NUM):
if self.cache_state[tempRSUId][
popularId] == 0 and self.rsu_residual_capcity[
tempRSUId] >= self.request_size[popularId]:
self.cache_state[tempRSUId][popularId] = 1
self.rsu_residual_capcity[tempRSUId] -= self.request_size[
popularId]
for tempCarId in range(CAR_NUM):
if self.car_cache_state[tempCarId][popularId] == 0 and self.car_residual_capcity[tempCarId] >= \
self.request_size[popularId]:
self.car_cache_state[tempCarId][popularId] = 1
self.car_residual_capcity[tempCarId] -= self.request_size[
popularId]
observation_ = np.concatenate([
self.rsu_residual_capcity, self.car_residual_capcity,
self.request_popularity
])
store_cost = (np.sum(self.rsu_capcity) + np.sum(self.car_capcity) -
np.sum(self.rsu_residual_capcity) -
np.sum(self.car_residual_capcity)) / (
np.sum(self.rsu_capcity) + np.sum(self.car_capcity))
tranTime_sum = 0
for i in range(REGION_NUM):
for j in range(REQUEST_NUM):
tempRSUId = self.region_rsu[i]
if self.cache_state[tempRSUId][j] == 1:
tranTime_sum += self.time2
else:
tranTime_sum += self.time4
trans_cost = tranTime_sum / (self.time4 * REQUEST_NUM * REGION_NUM)
cost = store_cost + trans_cost
#print(store_cost,trans_cost)
reward = -cost
self.region_request = self.requestGenerator.generateRegionRequestMatrix(
)
self.carmove_step()
return observation_, reward, currentRequestCount, hitCacheCount
def carmove_step(self):
for i in range(len(self.car_location)):
current_location = self.car_location[i]
tempint = random.randint(1, 100)
if tempint > 90:
candi_location = []
for j in range(len(self.region_neighbor[current_location])):
if current_location != j and self.region_neighbor[
current_location][j] == 1:
candi_location.append(j)
candi_location_choose = random.randint(0,
len(candi_location) - 1)
self.car_location[i] = candi_location[candi_location_choose]
def hitCacheCount(self):
requestCount = 0
cacheHitCount = 0
for i in range(len(self.region_request)):
for j in range(len(self.region_request[0])):
if self.region_request[i][j] == 1:
requestCount += 1
correspondRsuId = self.region_rsu[i]
if self.cache_state[correspondRsuId][j] == 1:
cacheHitCount += 1
else:
for car_id in range(len(self.car_location)):
if self.car_location[
car_id] == i and self.car_cache_state[
car_id][j] == 1:
cacheHitCount += 1
break
return requestCount, cacheHitCount
def calculatePopularity(self):
for i in range(REGION_NUM):
for j in range(REQUEST_NUM):
if self.region_request[i][j] == 1:
self.request_popularity[j] += 1