-
Notifications
You must be signed in to change notification settings - Fork 0
/
k_split_GCSQ_exactly.py
45 lines (39 loc) · 2.33 KB
/
k_split_GCSQ_exactly.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
import utils
import numpy as np
from algorithm import IterativeQuantumAlgorithmWithK
class k_split_GCSQ_exactly(IterativeQuantumAlgorithmWithK):
def __init__(self, seed, num_graph_sizes, k, solver="qbsolv", timeout=600, parallel=True):
super().__init__(seed=seed, num_graph_sizes=num_graph_sizes, solver=solver, timeout=timeout, k=k, parallel=parallel)
self.name = f"{self.k}_split_GCSQ_exactly_{self.solver}_{'parallel' if self.parallel else 'sequential'}"
def _get_qubo(self, coalition, edges):
Q = {}
# TODO: Find good value for penalty using penalty engineering
# sum of all the edges absolute values to use as a penalty value later
penalty = np.sum(np.abs(list(edges.values())))
# iterate over agents vertically (rows)
for i in range(len(coalition)):
for c in range(self.k):
# get number of logical qubit vertically (rows)
q_ic = i * self.k + c
# iterate over agents horizontally (columns)
for j in range(i + 1, len(coalition)):
# get number of logical qubit horizontally (columns)
q_jc = j * self.k + c
utils.add(Q, q_ic, q_ic, edges[(coalition[i], coalition[j])])
utils.add(Q, q_jc, q_jc, edges[(coalition[i], coalition[j])])
utils.add(Q, q_ic, q_jc, -2 * edges[(coalition[i], coalition[j])])
# add reward for putting agent in any coalition
utils.add(Q, q_ic, q_ic, -penalty)
for c2 in range(c + 1, self.k):
# get number of logical qubit horizontally (columns)
q_ic2 = i * self.k + c2
# add penalty for putting agent in two different coalitions at the same time (we don't do overlap here yet)
utils.add(Q, q_ic, q_ic2, 2 * penalty)
return Q, self.k * len(coalition)
def _get_coalitions_from_qubo_solution(self, coalition, solution):
# make a list of coalitions, with each coalition being a list with the numbers of the agents in these coalitions
coalitions = []
for c in range(self.k):
new_coalition = [coalition[i] for i in range(len(coalition)) if solution[i * self.k + c] == 1]
coalitions.append(new_coalition)
return coalitions