... literals.¹

It is easy to see that ${\cal M}(f)$ and $\hat{\xi}(f)$ are readily related. Specifically each constituent contains $k$ of the $\vert F\vert$ literals, corresponding to $2^{\vert F\vert-k}$ models.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... paper.²

We are aware of the negative publicity associated with the B&T problems and we do in fact handle more interesting problems with difficult reachability and uncertainty (e.g. Logistics and Rovers), but to simplify our discussion we choose this small problem.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... as:³

Note that $BS'$ may not be in clausal form after regression (especially when an action has multiple conditional effects).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... distance.⁴

Interaction between states captures the notion that actions performed to transition one state to the goal may interfere (negatively interact) or aid with (positively interact) transitioning other states to goals states.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... knowledge.⁵

This problem may be interesting to investigate in a multi-agent planning scenario, assuming no global communication (e.g. no radio dispatcher).

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

... heuristics⁶

We show $h_{s-RP}^{MG}$ with $POND$ .

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.