process tidyup-mdp AI-Planning#4

README.md

@@ -19,7 +19,6 @@ Planning under non-deterministic `oneof` actions was then first used in the 2006
 - `forest` and `forest-new` had `problem` type renamed to `sub-problem` to avoid a conflict with the `problem` keyword.
 - Missing requirements added to `st_blocksworld`, `st_first_responders`, `st_mapfdu`, `st_tires`, `st_faults`, `puffbot_dialogue_pddl`, `tidyup-mdp`, and `zenotravel`.
 - `puffbot_dialogue_pddl` had the problem-specific objects moved to domain constants, since they are used in actions.
-- `tidyup-mdp`: modified to fix some PDDL bugs: the `cup*` objects were moved from the problem files to constants in the domain, since they are mentioned explicitly in action preconditions.
 
 
 ## Probabilistic Interesting  - (Little and Thiebaux 2007)
@@ -91,6 +90,8 @@ These five come from the [FOND-SAT](https://github.com/tomsons22/FOND-SAT) repo/
 - `chain-of-rooms`: from [Constructing conditional plans by a theorem-prover](https://jair.org/index.php/jair/article/view/10230/), JAIR 10 (1999), 323–352, as part of the QBFPLAN planner.
 - `earth-observation`: ([J Aldinger & J Löhr, 2013)](https://gki.informatik.uni-freiburg.de/papers/aldinger-loehr-pcd2013.pdf)
 - `puffbot-dialog`: domain generating the dialogue policy for building dialogue agents in the health-care setting. Reported in [Teixeira et. al SAC'21 paper](https://dl.acm.org/doi/10.1145/3412841.3441942) and [PhD thesis](https://iris.unitn.it/retrieve/handle/11572/361402/595412/phd_unitn_santos_teixeira_milene.pdf).
+- `tidyup-mdp`: from MyND (as far as we can trace it back; see [here](https://bitbucket.org/robertmattmueller/mynd/src/master/data/fond-pddl/tidyup-mdp/)); with significant clean-up by Christian Muise.
+  - modified to fix some PDDL bugs: the `cup*` objects were moved from the problem files to constants in the domain, since they are mentioned explicitly in action preconditions.
 
 ### From strong domains: `st_xyz`