Is it curtains for causal encodings?
Results here seem paradoxical given all the attention partial-order planners got in conjunctive planning
- But there is really no paradox!
- Causal proofs have
- Flexibility to insert steps anywhere in the plan
- Ability to take causal relations into account
- The cost of this flexibility is that they are localied (longer) than state-based proofs
Neither flexibility is needed in solving SAT encodings
Useful for incremental planning/Plan Reuse
Useful for enforcing causal constraints