cbc-casper
cbc-casper copied to clipboard
Oracle Comparisons
Issue
Currently, we have no way of figuring out how efficient different safety oracles are in different situations. It would be nice to be able to compare these different oracles in different situations.
Proposed Implementation
Create an automated mechanism for testing oracles in different situations, and giving us useful information about their run-times, how high of a lower-bound they can recognize, etc.
Side note. networkX is great as it provides many built in algs and plotting tool out of the box, but it may be worth exploring building our own graph data structure from scratch that implements our more fine tuned clique searching algorithms w/ plotting and analytic hooks.