chap3part3chap4 - CS2710,ISSP2610 Chapter3,Part3...

Info iconThis preview shows pages 1–13. Sign up to view the full content.

View Full Document Right Arrow Icon
1 CS 2710, ISSP 2610 Chapter 3, Part 3 Heuristic Search Chapter 4 Local Search and Optimization
Background image of page 1

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
2 Beam Search Cheap, unpredictable search For problems with many solutions, it may be  worthwhile to discard unpromising paths Greedy best first search that keeps a fixed number  of nodes on the fringe
Background image of page 2
3 Beam Search def beamSearch(fringe,beamwidth):     while len(fringe) > 0:        cur = fringe[0]        fringe = fringe[1:]        if goalp(cur):  return cur        newnodes = makeNodes(cur, successors(cur))        for s in newnodes:           fringe = insertByH(s, fringe)        fringe = fringe[:beamwidth] return []
Background image of page 3

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
4 Beam Search Optimal?  Complete? Hardly! Space? O(b)  (generates the successors) Often useful
Background image of page 4
Creating Heuristics 5
Background image of page 5

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
6 Combining Heuristics If you have lots of heuristics and none dominates  the others and all are admissible… Use them all! H(n) = max(h1(n), …, hm(n))
Background image of page 6
Relaxed Heuristic Relaxed problem A problem with fewer restrictions on the actions The cost of an optimal solution to a relaxed problem  is an admissible heuristic for the original problem. 7
Background image of page 7

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
8 Relaxed Problems Exact solutions to different (relaxed) problems H1  (# of misplaced tiles)  is perfectly accurate if a  tile could move to any square H2  (sum of Manhattan distances)  is perfectly  accurate if a tile could move 1 square in any  direction
Background image of page 8
9 Relaxed Problems If problem is defined formally as a set of constraints,  relaxed problems can be generated automatically Absolver (Prieditis, 1993) Discovered a better heuristic for 8 puzzle and the first useful  heuristic for Rubik’s cube
Background image of page 9

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Systematic Relaxation Precondition List A conjunction of predicates that must hold true before the action can be applied Add List A list of predicates that are to be added to the description of the world-state as a  result of applying the action Delete List A list of predicates that are no longer true once the action is applied and should,  therefore, be deleted from the state description Primitive Predicates ON(x, y) : tile x is on cell y CLEAR(y) : cell y is clear of tiles ADJ(y, z) : cell y is adjacent to cell z 10
Background image of page 10
    Here is the full definition of s move for the n-puzzle Move(x, y, z): precondition list ON(x, y), CLEAR(z), ADJ(y, z) add list ON(x, z), CLEAR(y) delete list ON(x, y), CLEAR(z) 11
Background image of page 11

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
 
Background image of page 12
Image of page 13
This is the end of the preview. Sign up to access the rest of the document.

Page1 / 59

chap3part3chap4 - CS2710,ISSP2610 Chapter3,Part3...

This preview shows document pages 1 - 13. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online