def test_astar(self):
'algorithm = "astar"'
alg = GreedyElim(self.G, self.max_res, self.min_res, algorithm="astar")
alg.run()
self.assertEqual(alg.path, self.result_path)
self.assertEqual(alg.total_cost, self.total_cost)
self.assertTrue(all(alg.consumed_resources == self.consumed_resources))
def test_threshold(self):
'test threshold parameter'
alg = GreedyElim(self.G, self.max_res, self.min_res, threshold=100)
alg.run()
self.assertEqual(alg.path, ["Source", "A", "B", "Sink"])
self.assertEqual(alg.total_cost, 8)
self.assertTrue(all(alg.consumed_resources == [3, 4.3]))
def test_simple(self):
'algorithm = "simple" (default)'
alg = GreedyElim(self.G, self.max_res, self.min_res)
alg.run()
self.assertEqual(alg.path, self.result_path)
self.assertEqual(alg.total_cost, self.total_cost)
self.assertTrue(all(alg.consumed_resources == self.consumed_resources))
def test_time_limit(self):
'time limit parameter'
alg = GreedyElim(self.G, self.max_res, self.min_res, time_limit=0.001)
start = time()
alg.run()
self.assertTrue(time() - start <= 0.001)
self.assertEqual(alg.path, self.result_path)
self.assertEqual(alg.total_cost, self.total_cost)
self.assertTrue(all(alg.consumed_resources == self.consumed_resources))
def test_greedyelim(self):
alg = GreedyElim(self.G,
self.max_res,
self.min_res,
REF_callback=self.my_callback)
alg.run()
self.assertEqual(alg.path, self.result_path_heur)
self.assertEqual(alg.total_cost, self.total_cost_heur)
self.assertEqual(list(alg.consumed_resources),
self.consumed_resources_heur)
def test_astar_threshold(self):
'test threshold parameter'
alg = GreedyElim(self.G,
self.max_res,
self.min_res,
algorithm="astar",
threshold=100)
alg.run()
self.assertEqual(alg.path, self.result_path)
self.assertEqual(alg.total_cost, self.total_cost)
self.assertTrue(all(alg.consumed_resources == self.consumed_resources))
def test_time_limit_threshold(self):
'time limit and threshold parameters'
alg = GreedyElim(self.G,
self.max_res,
self.min_res,
time_limit=0.001,
threshold=100)
start = time()
alg.run()
self.assertTrue(time() - start <= 0.001)
self.assertEqual(alg.path, ["Source", "A", "B", "Sink"])
self.assertEqual(alg.total_cost, 8)
self.assertTrue(all(alg.consumed_resources == [3, 4.3]))
def test_astar_time_limit_threshold(self):
'time limit and threshold parameters'
alg = GreedyElim(self.G,
self.max_res,
self.min_res,
algorithm="astar",
time_limit=0.001,
threshold=100)
start = time()
alg.run()
self.assertTrue(time() - start <= 0.001)
self.assertEqual(alg.path, self.result_path)
self.assertEqual(alg.total_cost, self.total_cost)
self.assertTrue(all(alg.consumed_resources == self.consumed_resources))
def solve(self, time_limit):
"""
Solves the subproblem with cspy.
Resolves until:
1. heuristic algorithm gives a new route (column with -ve reduced cost);
2. exact algorithm gives a new route;
3. neither heuristic nor exact give a new route.
Note : time_limit has no effect for the moment
"""
if not self.run_subsolve:
return self.routes, False
self.formulate()
logger.debug("resources = {}".format(self.resources))
logger.debug("min res = {}".format(self.min_res))
logger.debug("max res = {}".format(self.max_res))
more_routes = False
while True:
if self.exact:
logger.debug("solving with bidirectional")
self.alg = BiDirectional(
self.sub_G,
self.max_res,
self.min_res,
direction="both",
method="generated",
REF_forward=self.get_REF("forward"),
REF_backward=self.get_REF("backward"),
REF_join=self.get_REF("join"),
)
else:
logger.debug("solving with greedyelim")
self.alg = GreedyElim(
self.sub_G,
self.max_res,
self.min_res,
REF=self.get_REF("forward"),
max_depth=40,
)
self.alg.run()
logger.debug("subproblem")
logger.debug("cost = %s" % self.alg.total_cost)
logger.debug("resources = %s" % self.alg.consumed_resources)
if self.alg.total_cost < -(10**-3):
more_routes = True
self.add_new_route()
logger.debug("new route %s" % self.alg.path)
logger.debug("reduced cost = %s" % self.alg.total_cost)
logger.debug("real cost = %s" % self.total_cost)
break
# If not already solved exactly
elif not self.exact:
# Solve exactly from here on
self.exact = True
# Solved heuristically and exactly and no more routes
else:
break
return self.routes, more_routes
class SubProblemCSPY(SubProblemBase):
"""
Solves the sub problem for the column generation procedure with cspy;
attemps to find routes with negative reduced cost.
Inherits problem parameters from `SubproblemBase`
"""
def __init__(self, *args, exact):
"""Initializes resources."""
# Pass arguments to base
super(SubProblemCSPY, self).__init__(*args)
# Resource names
self.alg = None
self.resources = [
"stops/mono",
"load",
"time",
"time windows",
"collect",
"deliver",
]
self.exact = exact
# Set number of resources as attribute of graph
self.sub_G.graph["n_res"] = len(self.resources)
# Default lower and upper bounds
self.min_res = [0] * len(self.resources)
# Add upper bounds for mono, stops, load and time, and time windows
total_demand = sum(
[self.sub_G.nodes[v]["demand"] for v in self.sub_G.nodes()])
self.max_res = [
len(self.sub_G.nodes()), # stop/mono
total_demand, # load
sum([
self.sub_G.edges[u, v]["time"] for u, v in self.sub_G.edges()
]), # time
1, # time windows
total_demand, # pickup
total_demand, # deliver
]
# Initialize cspy edge attributes
for edge in self.sub_G.edges(data=True):
edge[2]["res_cost"] = zeros(len(self.resources))
# Initialize max feasible arrival time
self.T = 0
# @profile
def solve(self, time_limit):
"""
Solves the subproblem with cspy.
Resolves until:
1. heuristic algorithm gives a new route (column with -ve reduced cost);
2. exact algorithm gives a new route;
3. neither heuristic nor exact give a new route.
Note : time_limit has no effect for the moment
"""
if not self.run_subsolve:
return self.routes, False
self.formulate()
logger.debug("resources = {}".format(self.resources))
logger.debug("min res = {}".format(self.min_res))
logger.debug("max res = {}".format(self.max_res))
more_routes = False
while True:
if self.exact:
logger.debug("solving with bidirectional")
self.alg = BiDirectional(
self.sub_G,
self.max_res,
self.min_res,
direction="both",
method="generated",
REF_forward=self.get_REF("forward"),
REF_backward=self.get_REF("backward"),
REF_join=self.get_REF("join"),
)
else:
logger.debug("solving with greedyelim")
self.alg = GreedyElim(
self.sub_G,
self.max_res,
self.min_res,
REF=self.get_REF("forward"),
max_depth=40,
)
self.alg.run()
logger.debug("subproblem")
logger.debug("cost = %s" % self.alg.total_cost)
logger.debug("resources = %s" % self.alg.consumed_resources)
if self.alg.total_cost < -(10**-3):
more_routes = True
self.add_new_route()
logger.debug("new route %s" % self.alg.path)
logger.debug("reduced cost = %s" % self.alg.total_cost)
logger.debug("real cost = %s" % self.total_cost)
break
# If not already solved exactly
elif not self.exact:
# Solve exactly from here on
self.exact = True
# Solved heuristically and exactly and no more routes
else:
break
return self.routes, more_routes
def formulate(self):
"""Updates max_res depending on which contraints are active."""
# Problem specific constraints
if self.num_stops:
self.add_max_stops()
else:
self.add_monotone()
if self.load_capacity:
self.add_max_load()
if self.duration:
self.add_max_duration()
if self.time_windows:
if not self.duration:
# Update upper bound for duration
self.max_res[2] = 1 + self.sub_G.nodes["Sink"]["upper"]
# Time windows feasibility
self.max_res[3] = 0
# Maximum feasible arrival time
self.T = max([
self.sub_G.nodes[v]["upper"] +
self.sub_G.nodes[v]["service_time"] +
self.sub_G.edges[v, "Sink"]["time"]
for v in self.sub_G.predecessors("Sink")
])
if self.load_capacity and self.distribution_collection:
self.max_res[4] = self.load_capacity[self.vehicle_type]
self.max_res[5] = self.load_capacity[self.vehicle_type]
def add_new_route(self):
"""Create new route as DiGraph and add to pool of columns"""
route_id = len(self.routes) + 1
new_route = DiGraph(name=route_id)
add_path(new_route, self.alg.path)
self.total_cost = 0
for (i, j) in new_route.edges():
edge_cost = self.sub_G.edges[i, j]["cost"][self.vehicle_type]
self.total_cost += edge_cost
new_route.edges[i, j]["cost"] = edge_cost
if i != "Source":
self.routes_with_node[i].append(new_route)
new_route.graph["cost"] = self.total_cost
new_route.graph["vehicle_type"] = self.vehicle_type
self.routes.append(new_route)
def add_max_stops(self):
"""Updates maximum number of stops."""
# Change label
self.resources[0] = "stops"
# The Sink does not count (hence + 1)
self.max_res[0] = self.num_stops + 1
for (i, j) in self.sub_G.edges():
self.sub_G.edges[i, j]["res_cost"][0] = 1
def add_monotone(self):
"""Updates monotone resource."""
# Change label
self.resources[0] = "mono"
for (i, j) in self.sub_G.edges():
self.sub_G.edges[i, j]["res_cost"][0] = 1
def add_max_load(self):
"""Updates maximum load."""
self.max_res[1] = self.load_capacity[self.vehicle_type]
for (i, j) in self.sub_G.edges():
demand_head_node = self.sub_G.nodes[j]["demand"]
self.sub_G.edges[i, j]["res_cost"][1] = demand_head_node
def add_max_duration(self):
"""Updates maximum travel time."""
self.max_res[2] = self.duration
for (i, j) in self.sub_G.edges():
travel_time = self.sub_G.edges[i, j]["time"]
self.sub_G.edges[i, j]["res_cost"][2] = travel_time
def get_REF(self, type_):
"""
Returns custom REFs if time, time windows, and/or distribution collection.
Based on Righini and Salani (2006).
"""
if self.time_windows or self.distribution_collection:
# Use custom REF
if type_ == "forward":
return self.REF_forward
elif type_ == "backward":
return self.REF_backward
elif type_ == "join":
return self.REF_join
else:
# Use default
return
def REF_forward(self, cumulative_res, edge):
"""
Resource extension for forward paths.
"""
new_res = array(cumulative_res)
# extract data
i, j = edge[0:2]
# stops/monotone resource
new_res[0] += 1
# load
new_res[1] += self.sub_G.nodes[j]["demand"]
# time
service_time = self.sub_G.nodes[i]["service_time"]
travel_time = self.sub_G.edges[i, j]["time"]
a_j = self.sub_G.nodes[j]["lower"]
b_j = self.sub_G.nodes[j]["upper"]
new_res[2] = max(new_res[2] + service_time + travel_time, a_j)
# time-window feasibility resource
if not self.time_windows or new_res[2] <= b_j:
new_res[3] = 0
else:
new_res[3] = 1
if self.distribution_collection:
# Pickup
new_res[4] += self.sub_G.nodes[j]["collect"]
# Delivery
new_res[5] = max(new_res[5] + self.sub_G.nodes[j]["demand"],
new_res[4])
return new_res
def REF_backward(self, cumulative_res, edge):
"""
Resource extension for backward paths.
"""
new_res = array(cumulative_res)
i, j = edge[0:2]
# monotone resource
new_res[0] -= 1
# load
new_res[1] += self.sub_G.nodes[i]["demand"]
# Get relevant service times (thetas) and travel time
theta_i = self.sub_G.nodes[i]["service_time"]
theta_j = self.sub_G.nodes[j]["service_time"]
travel_time = self.sub_G.edges[i, j]["time"]
# Lower time windows
a_i = self.sub_G.nodes[i]["lower"]
# Upper time windows
b_i = self.sub_G.nodes[i]["upper"]
new_res[2] = max(new_res[2] + theta_j + travel_time,
self.T - b_i - theta_i)
# time-window feasibility
if not self.time_windows or new_res[2] <= self.T - a_i - theta_i:
new_res[3] = 0
else:
new_res[3] = 1
if self.distribution_collection:
# Delivery
new_res[5] += new_res[5] + self.sub_G.nodes[i]["demand"]
# Pickup
new_res[4] = max(new_res[5],
new_res[4] + self.sub_G.nodes[i]["collect"])
return new_res
def REF_join(self, fwd_res, bwd_res, edge):
"""
Appropriate joining of forward and backward resources.
"""
final_res = zeros(len(self.resources))
i, j = edge[0:2]
# Get relevant service times (thetas) and travel time
theta_i = self.sub_G.nodes[i]["service_time"]
theta_j = self.sub_G.nodes[j]["service_time"]
travel_time = self.sub_G.edges[i, j]["time"]
# Invert monotone resource
bwd_res[0] = self.max_res[0] - bwd_res[0]
# Fill in final res
# Monotone / stops
final_res[0] = fwd_res[0] + bwd_res[0] + 1
# Load
final_res[1] = fwd_res[1] + bwd_res[1]
# time
final_res[
2] = fwd_res[2] + theta_i + travel_time + theta_j + bwd_res[2]
# Time windows
if not self.time_windows or final_res[2] <= self.T:
final_res[3] = fwd_res[3] + bwd_res[3]
else:
final_res[3] = 1
if self.distribution_collection:
final_res[4] = fwd_res[4] + bwd_res[4]
final_res[5] = fwd_res[5] + bwd_res[5]
return array(final_res)
def test_time_limit_raises(self):
'Time limit of 0 raises an exception'
alg = GreedyElim(self.G, self.max_res, self.min_res, time_limit=0)
self.assertRaises(Exception, alg.run)