
78 Chapter 4 State-Space Planning
loc1
crane1
p1
q1 q2 q3
p2
c1
c3
c3
c2
c1
c2
s
0
= {in(c3,p1), top(c3,p1), in(c1,p1), on(c3,c1)
on(c1,pallet), in(c2,p2), top(c2,p2),
on(c2,pallet), top(pallet,q1), top(pallet,q2)
top(pallet,q3), empty(crane1)}
g = {on(c1,c2)
on(c2,c3)}
Figure 4.5 A DWR version of the Sussman anomaly.
We can get the shortest plan for both goals by inserting the second plan between the
second and third actions of the first plan.
Observations such as these led to the development of a technique called plan-space
planning, in which the planning system searches through a space whose nodes are
partial plans rather than states of the world, and a partial ...