add max_weight_cycle helper function #1283
Conversation
Codecov Report
@@ Coverage Diff @@
## master #1283 +/- ##
=======================================
Coverage 98.14% 98.14%
=======================================
Files 147 147
Lines 11333 11344 +11
=======================================
+ Hits 11123 11134 +11
Misses 210 210
Continue to review full report at Codecov.
|
| @@ -19,6 +19,7 @@ | |||
| import networkx as nx | |||
| import numpy as np | |||
| import pennylane as qml | |||
| from pennylane.vqe import Hamiltonian | |||
There was a problem hiding this comment.
Doing this prevents problems with importing into PL.
| def find_simple_cycle(list_of_edges): | ||
| """Returns True if list_of_edges contains a permutation corresponding to a simple cycle""" | ||
| permutations = list(itertools.permutations(list_of_edges)) | ||
|
|
||
| for edges in permutations: | ||
| if edges[0][0] != edges[-1][-1]: # check first node is equal to last node | ||
| continue | ||
| all_nodes = [] | ||
| for edge in edges: | ||
| for n in edge: | ||
| all_nodes.append(n) | ||
| inner_nodes = all_nodes[ | ||
| 1:-1 | ||
| ] # find all nodes in all edges excluding the first and last nodes | ||
| nodes_out = [ | ||
| inner_nodes[i] for i in range(len(inner_nodes)) if i % 2 == 0 | ||
| ] # find the nodes each edge is leaving | ||
| node_in = [ | ||
| inner_nodes[i] for i in range(len(inner_nodes)) if i % 2 != 0 | ||
| ] # find the nodes each edge is entering | ||
| if nodes_out == node_in and ( | ||
| len([all_nodes[0]] + nodes_out) == len(set([all_nodes[0]] + nodes_out)) | ||
| ): # check that each edge connect to the next via a common node and that no node is crossed more than once | ||
| return True |
There was a problem hiding this comment.
Does this function allow for the (valid) case where two disconnected cycles are contained? I guess this case isn't possible anyway with 3 nodes 🤔
There was a problem hiding this comment.
I don't think it will since it checks the 1st and last nodes are equal, for example this disconnected pair of 2 cycles would not be classed as a valid case: disconnected = ((2,1), (1,2), (3,4), (4,3)) since
disconnected[0][0]!=disconnected[-1][-1]
But as you've pointed out, it will pick up all valid cycles for the purposes of this test.
Helper functions for the UI exist for each QAOA solution. This adds a helper function for the maximum weighted cycle functionality.
The penealty term size in the uncontrained approach is fixed. A user would have to access the penalty terms through
qaoa.cycleto control this.