slwu89
commented
2 years ago Right now we can't carry over matches from t-1 to t because even though we can update them with the new graphs using composition of partial morphisms we won't be able to search for new matches. Or we can, but we'll get duplicates of the ones we're carrying over, so we still need to wait for a way to search into restricted parts of the ACSet.
One way to improve the simulation would be to cache identified rule matches which did not fire from previous time steps (after checking if they have been made invalid). Then new matching should only look at the updated parts of state. For example, if last time we had 4 identified recovery events which didn't fire, keep them, and on this time step for finding more enabled recovery events, only look at those who have newly arrived in R.