Infomation Lower Bound Search #218
Conversation
| @@ -0,0 +1,150 @@ | |||
|
|
|||
| #include "experimental/beacon_sim/information_lower_bound_search.hh" | |||
There was a problem hiding this comment.
need to get rid of debugging prints
| env.road_map, env.start_idx, env.end_idx, START_INFO, GOAL_INFO_LOWER_BOUND, env.rev_prop); | ||
|
|
||
| // Verification | ||
| EXPECT_TRUE(false); |
| env.road_map, env.start_idx, env.end_idx, START_INFO, GOAL_INFO_LOWER_BOUND, env.rev_prop); | ||
|
|
||
| // Verification | ||
| EXPECT_TRUE(false); |
| const LowerBoundReversePropagator &propagator); | ||
|
|
||
| namespace detail { | ||
| struct InProgressPath { |
There was a problem hiding this comment.
Is there a case where InformationLowerBoundResult wouldn't exactly match InProgressPath?
There was a problem hiding this comment.
InProgressPath is meant to be an internal data structure that I'm only exposing for the sake of unit testing. I did want to separate the internal representation from the external API. I can easily imagine returning additional details about the search like the number of edges traversed, or the propagated bounds at each node as a part of the result.
|
|
||
| struct MergeResult { | ||
| bool should_merge; | ||
| std::vector<int> to_boot; |
There was a problem hiding this comment.
What is to_boot? A short comment could help
There was a problem hiding this comment.
Each node has a list of paths from that node to the goal node, along with the information lower bound required for that path and the cost of that path. Given a new path to a node, should_merge() returns:
a) If that new path should be added to the list of paths maintained by that node, and
b) If that new path should be added, which paths, if any, should be dropped from the list of paths maintained by the node. This is represented by the vector of indices that should be dropped or booted, hence to_boot.
I'll add some comments :).
| continue; | ||
| } | ||
|
|
||
| // remove any paths that have been dominated |
There was a problem hiding this comment.
Does this remove the path from all nodes along that partial path? Some additional comments could help quick understanding of this complicated bit of code
There was a problem hiding this comment.
Removal from previous nodes doesn't happen here. I'll add comments adding more details as I'm almost certainly going to revisit this piece of code.
No description provided.