1. Search Model 2
Using a tree we don’t need to keep information about previous states in the search control, rather this data is kept in the search model itself
Tyson Kendon © 2007

2. Search Model 2 s0= (0,-) (1, 2, 3, 4, 5, 6, 7, 8) (1,1)
Tyson Kendon © 2007

3. Search Model 2  (0,-) (1, 2, 3, 4, 5, 6, 7, 8) (1,1) (7,d) (0,-)
(1,u)
(0,-)
(1, 2, 3, 4, 5, 6, 8)
(1,2)
(2, 3, 4, 5, 6, 7, 8)
(1,2)
(3,r)
(0,-)
(6r,-)
(0,-)
(1, 2, 4, 5, 6, 7, 8)
(1,2)
(1, 2, 3, 4, 5, 7, 8)
(1,2)
Tyson Kendon © 2007

4. Search Model 2   Tyson Kendon © 2007 (0,-) (1, 2, 3, 4, 5, 6, 7, 8)
(1,1) 
(7,u)
(0,-)
(1,u)
(0,-)
(1, 2, 3, 4, 5, 6, 8)
(1,2)
(2, 3, 4, 5, 6, 7, 8)
(1,2)
(3,r)
(0,-)
(6,r)
(0,-)
(1,u)
(4,u)
(0,-)
(1, 2, 4, 5, 6, 7, 8)
(1,2)
(1, 2, 3, 4, 5, 7, 8)
(1,2)
(1, 2, 3, 5, 7, 8)
(2,1)
Tyson Kendon © 2007

5. Search Model 2    Tyson Kendon © 2007 (0,-)
(1, 2, 3, 4, 5, 6, 7, 8)
(1,1) 
(7,u)
(0,-)
(1,u)
(0,-)
(1, 2, 3, 4, 5, 6, 8)
(1,2)
(2, 3, 4, 5, 6, 7, 8)
(1,2)
(3,r)
(0,-)
(6,r)
(0,-) 
(1,u)
(4,u)
(0,-)
(1, 2, 4, 5, 6, 7, 8)
(1,2)
(1, 2, 3, 4, 5, 7, 8)
(1,2)
(2, 3, 5, 6, 7, 8)
(2,1)
(1,u)
(4,u)
(0,-)
(6,d)
Tyson Kendon © 2007
(2, 3, 5, 7, 8)
(1,3)

6. Search Model 2    Tyson Kendon © 2007 (1,u) (0,-)
(2, 3, 4, 5, 6, 7, 8)
(1,2) 
(1,u)
(4,u)
(0,-)
(2, 3, 5, 6, 7, 8)
(2,1) 
(1,u)
(4,u)
(0,-)
(6,d)
(2, 3, 5, 7, 8)
(1,3)
(1,u)
(4,u)
(8,u)
(0,-)
(6,d)
Tyson Kendon © 2007
(2, 3, 5, 7)
(2,2)

7. Search Model 2    Tyson Kendon © 2007 (1,u) (4,u) (0,-)
(2, 3, 5, 6, 7, 8)
(2,1) 
(1,u)
(4,u)
(0,-)
(6,d)
(2, 3, 5, 7, 8)
(1,3) 
(1,u)
(4,u)
(8,u)
(0,-)
(6,d)
(2, 3, 5, 7)
(2,2)
(1,u)
(4,u)
(8,u)
(0,-)
(6,d)
(5,d)
Tyson Kendon © 2007
(2, 3, 7)
(1,4)

8. Search Model 2    Tyson Kendon © 2007 (1,u) (4,u) (0,-) (6,d)
(2, 3, 5, 7, 8)
(1,3) 
(1,u)
(4,u)
(8,u)
(0,-)
(6,d)
(2, 3, 5, 7)
(2,2) 
(1,u)
(4,u)
(8,u)
(0,-)
(6,d)
(5,d)
(2, 3, 7)
(1,4)
(1,u)
(4,u)
(8,u)
(3,u)
(6,d)
(5,d)
(0,-)
Tyson Kendon © 2007
(2, 7)
(2,3)

9. Search Model 2    Tyson Kendon © 2007 (1,u) (4,u) (8,u) (0,-) (6,d)
(2, 3, 5, 7)
(2,2) 
(1,u)
(4,u)
(8,u)
(0,-)
(6,d)
(5,d)
(2, 3, 7)
(1,4) 
(1,u)
(4,u)
(8,u)
(3,u)
(6,d)
(5,d)
(0,-)
(2, 7)
(2,3)
(1,u)
(4,u)
(8,u)
(3,u)
(6,d)
(5,d)
(2,d)
(0,-)
Tyson Kendon © 2007
(7)
(2,4)

10. Search Model 2    Tyson Kendon © 2007 (1,u) (4,u) (8,u) (0,-) (6,d)
(2, 3, 7)
(1,4) 
(1,u)
(4,u)
(8,u)
(3,u)
(6,d)
(5,d)
(0,-)
(2, 7)
(2,3) 
(1,u)
(4,u)
(8,u)
(3,u)
(6,d)
(5,d)
(2,d)
(0,-)
(7)
(2,4)
(1,u)
(4,u)
(8,u)
(3,u)
(6,d)
(5,d)
(2,d)
(7,d)
Tyson Kendon © 2007 ()
(-,-)

11. Tyson Kendon © 2007

12. Search Model 2
Gpuzzle2(s9) = yes, since s9 has a node where Rest is empty
This search control worked very well for the problem example, on another example it may have generated many more nodes, vice versa another control might have been less efficient on this example
Tyson Kendon © 2007

13. Notes
Change a node to FAIL if you cannot extend the matrix to any successors
Tyson Kendon © 2007

14. Notes  Tyson Kendon © 2007 (0,-) (1, 2, 3, 4, 5, 6, 7, 8) (1,1) (7,u)
(1, 2, 3, 4, 5, 6, 8)
(1,2)
(1,u)
(0,-)
(3,r)
(0,-)
(6,r)
(0,-)
(2, 3, 4, 5, 6, 7, 8)
(1,2)
(1, 2, 4, 5, 6, 7, 8)
(1,2)
(1, 2, 3, 4, 5, 7, 8)
(1,2)
Tyson Kendon © 2007

15. Notes (-,-) FAIL   Tyson Kendon © 2007 (0,-)
(1, 2, 3, 4, 5, 6, 7, 8)
(1,1)
(7,u)
(0,-)
(1, 2, 3, 4, 5, 6, 8)
(1,2) 
(1,u)
(0,-)
(6,r)
(0,-)
(-,-)
FAIL
(2, 3, 4, 5, 6, 7, 8)
(1,2)
(1, 2, 3, 4, 5, 7, 8)
(1,2)
Tyson Kendon © 2007