-
Notifications
You must be signed in to change notification settings - Fork 0
/
Solver.py
285 lines (228 loc) · 10.3 KB
/
Solver.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
import random, copy
import AdaptiveTuning as tune
from Model import *
from Utils import *
from Testing import *
from Optimization import *
class Solution:
"""Represents found solution
Attributes:
- profit: Profit number
- routes: List containing vehicle routes
"""
def __init__(self):
self.profit = 0.0
self.duration = 0.0
self.routes = []
class CustomerInsertion(object):
"""Represents a node insertion in a route
To be used for customer nodes
Attributes:
- customer: Customer `Node` for insertion
- route: `Route` for customer to be inserted
- profit: Profit gained from insertion
"""
def __init__(self):
self.customer = None
self.route = None
self.profit = 0
class CustomerInsertionAllPositions(object):
"""Represents a node insertion in a specific
position of a route
To be used for customer nodes
Attributes:
- customer: Customer `Node` for insertion
- route: `Route` for customer to be inserted
- insertionPosition: Position number for insertion
- profit: Profit gained from insertion
"""
def __init__(self):
self.customer = None
self.route = None
self.insertionPosition = None
self.profit = 0
class SavingsObject():
"""Represents the distance saved is two nodes are merged
To be used for Clarke-Wright
Attributes:
- i: start node
- j: end node
- distanceSaved: the total distance saved by the merge
"""
def __init__(self):
self.i = None
self.j = None
self.distanceSaved = 0
def __init__(self, i, j, distanceSaved):
self.i = i
self.j = j
self.distanceSaved = distanceSaved
class RandomCandidate:
def __init__(self, customer: Node, trialProfit: float, \
route: Route, insertionPosition: int):
self.customer = customer
self.trialProfit = trialProfit
self.route = route
self.insertionPosition = insertionPosition
class Solver:
"""Class to solve built problem model
Attributes:
- allNodes: List of all model nodes
- customers: List of all nodes representing customers
- depot: Depot node
- distanceMatrix: List representing a matrix of all node distances
- capacity: Max capacity of vehicles
- duration: Max available time for customer service
- vehicles: Available vehicles
- sol: current `Solution`
- overallBestSol: Overall best `Solution`
- rcl_size: Number of elements to be used in restricted candidate list
"""
def __init__(self, m):
self.allNodes: list[Node] = m.allNodes
self.customers: list[Node] = m.customers
self.depot: Node = m.allNodes[0]
self.distanceMatrix = m.distances
self.capacity = int(m.max_capacity)
self.duration = int(m.max_duration)
self.vehicles = int(m.vehicles)
self.constraints = {"capacity": self.capacity, "duration": self.duration, "vehicles": self.vehicles}
self.sol: Solution = None
self.overallBestSol: Solution = None
self.rcl_size = tune.rclSize
def solve(self):
for seed in range(10, 60, 10):
sol = self.MinimumInsertions(itr=seed, foundSolution=None)
if self.overallBestSol == None or self.overallBestSol.profit < sol.profit:
self.overallBestSol = copy.copy(sol)
self.overallBestSol.duration = CalculateTotalDuration(self.distanceMatrix, self.overallBestSol)
print("profit before vns")
print(self.overallBestSol.profit)
self.overallBestSol = VNS(self.overallBestSol, 2, self.distanceMatrix)
self.overallBestSol.duration = CalculateTotalDuration(self.distanceMatrix, self.overallBestSol)
for seed in range(10, 60, 10):
sol = self.MinimumInsertions(itr=seed, foundSolution=self.overallBestSol)
if self.overallBestSol == None or self.overallBestSol.profit < sol.profit:
self.overallBestSol = copy.copy(sol)
print("profit after vns")
return self.overallBestSol
def NearestNeighbor(self, itr=30) -> Solution:
solution = Solution()
solution.routes.append(Route(self.depot, self.capacity, self.duration))
pool = set(self.customers)
vehiclesUsed = 1
while vehiclesUsed <= 6:
rt = solution.routes[-1]
insertCust = self.FindBestNN(pool, rt, itr)
if insertCust:
# before the second occurence of depot
insIndex = len(rt.sequenceOfNodes) - 1
rt.sequenceOfNodes.insert(insIndex, insertCust)
rt.profit += insertCust.profit
rt.travelled = CalculateTravelledTime(self.distanceMatrix, rt)
rt.load += insertCust.demand
pool.remove(insertCust)
else:
solution.profit += rt.profit
solution.duration += rt.travelled
vehiclesUsed += 1
if len(solution.routes) < 6:
solution.routes.append(Route(self.depot, self.capacity, self.duration))
return solution
def FindBestNN(self, pool: list[Node], route: Route, itr) -> Node:
rng = random.Random(itr)
rcl: list[RandomCandidate] = []
for cust in pool:
if route.load + cust.demand <= route.capacity and \
AppendNodeDuration(self.distanceMatrix, route, cust) \
+ route.travelled <= route.duration:
trialProfit = math.pow(cust.profit, tune.nnNumerator) / \
math.pow(AppendNodeDuration(self.distanceMatrix, route, cust), tune.nnDenominator)
candidate = RandomCandidate(cust, trialProfit, route, route.sequenceOfNodes[-1])
# Update rcl list
if len(rcl) <= self.rcl_size:
rcl.append(candidate)
rcl.sort(key=lambda x: x.trialProfit)
elif candidate.trialProfit > rcl[0].trialProfit - tune.precision:
rcl.pop(0)
rcl.append(candidate)
rcl.sort(key=lambda x: x.trialProfit)
if len(rcl) == 0:
return # No fit candidates left
# Choose a candidate randomly
candidateIndex = rng.randint(0, len(rcl) - 1)
return rcl[candidateIndex].customer
def MinimumInsertions(self, itr=30, foundSolution: Solution = None) -> Solution:
"""Implements insertions algorithm
Can both build a solution from scratch, as well as improve a given solution.
Args:
itr (`int`, optional): Seed to use in rng. Defaults to 30.
foundSolution (`Solution`, optional): Already found solution. Defaults to None.
Returns:
Solution: Solution found with algorithm
"""
pool = set(self.customers)
solution = Solution()
if foundSolution:
sequences = list(map(lambda x: x.sequenceOfNodes, foundSolution.routes))
routedCustomers = set().union(*sequences)
pool = pool.difference(routedCustomers)
solution.routes.extend(foundSolution.routes)
else:
solution.routes.append(Route(self.depot, self.capacity, self.duration))
termination = False
while not termination:
candidate = self.FindBestInsertion(pool, solution.routes, itr)
if candidate: # Found insertion
insertCust = candidate.customer
rt = candidate.route
pos = candidate.insertionPosition
# Apply insertion
rt.sequenceOfNodes.insert(pos, insertCust)
rt.load += insertCust.demand
rt.travelled = CalculateTravelledTime(self.distanceMatrix, rt)
rt.profit += insertCust.profit
pool.remove(insertCust)
else: # No possible insertion
if len(solution.routes) < 6:
solution.routes.append(Route(self.depot, self.capacity, self.duration))
else:
termination = True
for r in solution.routes:
solution.duration += r.travelled
solution.profit += r.profit
return solution
def FindBestInsertion(self, pool: set[Node], routes: list[Route], itr) -> RandomCandidate:
rng = random.Random(itr)
rcl: list[RandomCandidate] = []
for cust in pool:
for route in routes:
# Check capacity constraint & PART of time constraint
if route.load + cust.demand <= route.capacity and \
cust.service_time + route.travelled <= route.duration:
for pos in range(len(route.sequenceOfNodes) - 1):
A: Node = route.sequenceOfNodes[pos]
B: Node = route.sequenceOfNodes[pos + 1]
costAdded = self.distanceMatrix[A.id][cust.id] + \
self.distanceMatrix[cust.id][B.id] + cust.service_time
costRemoved = self.distanceMatrix[A.id][B.id]
Dc = costAdded - costRemoved
# Check time constraint fully
if route.travelled + Dc > route.duration:
continue
Dp = cust.profit
trialProfit = math.pow(Dp, tune.minInsNumerator) / math.pow(Dc, tune.minInsDenominator)
candidate = RandomCandidate(cust, trialProfit, route, pos + 1)
# Update rcl list
if len(rcl) <= self.rcl_size:
rcl.append(candidate)
rcl.sort(key=lambda x: x.trialProfit)
elif candidate.trialProfit > rcl[0].trialProfit - tune.precision:
rcl.pop(0)
rcl.append(candidate)
rcl.sort(key=lambda x: x.trialProfit)
if len(rcl) == 0:
return None # No fit candidates left
# Choose a candidate randomly
candidateIndex = rng.randint(0, len(rcl) - 1)
return rcl[candidateIndex]