FOND solver based on SAT (https://arxiv.org/pdf/1806.09455.pdf)
Tomas Geffner and Hector Geffner
- F-domains/ contains the FOND domains used
- src/ contains the code for the solver, and a pre-compiled version of Minisat
Example usage (basic)
From the src/ directory run the following command:
python main.py ../F-domains/islands/domain.pddl ../F-domains/islands/p03.pddl
This would run the solver for the task 03 of the islands domain. The path to the domain and the task must be included.
Other options (arguments when calling)
- -time_limit: to set the time limit
- -mem_limit: to set the memory limit
- -name_temp: to set the name for the temporary files (eg. where the SAT formula is stored). If running the planner for many tasks in parallel use different temp names to avoid erroneous replacement of files
- -strong: to set the planner to look for strong policies
- -inc: to set the increment in the size of the controller each step. By default the planner looks for a solution of size 2, if it does not find one it looks for a solution of size 3, and so on. If inc is set to i, the planner looks for a solution of size 2, if it does not find one it looks for a solution of size 2+i, and so on
- -policy: to set whether the final policy (if found) should be displayed
Interpreting the policy
The policy displayed has 4 sections:
- Atom (CS): For each Controller State (CS) it prints out which atoms are true.
- (CS, Action with arguments): For each CS it prints what actions are applied in it.
- (CS, Action name, CS): For each CS it prints the action applied in that state (without arguments, for action with arguments check second section) and successor CS.
- (CS, CS): (cs1, cs2) means that the controller can go from cs1 to cs2. In other words, the action applied in cs1 may lead to cs2.
For easiness to use, this includes a pre-compiled version of Minisat. This pre-compiled version does not allow the use of time/memory limits. If you want to try another SAT solver (or use time/memory contraints), the following parts of the code should be modified:
- src/main.py: comment line 117.
- src/main.py: uncomment line 118 and adapt it to the corresponding SAT solver.
- src/CNF.py: line 200 corresponds to the function parseOutput(...) which reads the file output by the SAT solver. This function should be modified to parse the output of your SAT solver (works for versions of Minisat).
- RECOMMENDATION: install a version of Minisat from http://minisat.se/, comment line 117 of main.py and uncomment line 118. Using another version of Minisat will allow the use of time/memory constraints (newer version is also faster), and does not require the modification of parseOutput(...). The results shown in the paper were obtained using a newer version of Minisat, not the pre-compiled one.
Dual FOND planning
The paper talks about what we call Dual FOND planning. Dual FOND problems are those in which some actions are fair and some are unfair. To set some action (or actions) as unfair, add _unfair_ as the last part of the action name in the pddl file. The planner will then set this action as unfair.
The solver uses the parser from PRP (https://bitbucket.org/haz/planner-for-relevant-policies/wiki/Home)