1. Searching for Solutions Artificial Intelligence CMSC 25000 January 15, 2008

2. Agenda Search – Motivation –Problem-solving agents –Rigorous problem definitions Blind exhaustive search: –Breadth-first search –Uniform search –Depth-first search –Iterative deepening search Search analysis –Computational cost, limitations

3. Problem-Solving Agents Goal-based agents –Identify goal, sequence of actions that satisfy Goal: set of satisfying world states –Precise specification of what to achieve Problem formulation: –Identify states and actions to consider in achieving goal –Given a set of actions, consider sequence of actions leading to a state with some value Search: Process of looking for sequence –Problem -> action sequence solution

4. Agent Environment Specification ● Dimensions – Fully observable vs partially observable: Fully – Deterministic vs stochastic: Deterministic – Static vs dynamic: Static – Discrete vs continuous: Discrete ● Issues?

5. Closer to Reality Sensorless agents (conformant problems) –Replace state with “belief state” Multiple physical states, successors:sets of successors Partial observability (contigency problems) –Solution is tree, branch chosen based on percepts

6. Example: vacuum world Single-state, start in #5. Solution?

7. Example: vacuum world Single-state, start in #5. Solution? [Right, Vacuum] Sensorless, start in {1,2,3,4,5,6,7,8} e.g., Right goes to {2,4,6,8} Solution?

8. Example: vacuum world Sensorless, start in {1,2,3,4,5,6,7,8} e.g., Right goes to {2,4,6,8} Solution? [Right,Vacuum,Left,Vacuum] Contingency –Nondeterministic: Vacuum may dirty a clean carpet –Partially observable: location, dirt at current location. –Percept: [L, Clean], i.e., start in #5 or #7 Solution?

9. Example: vacuum world Sensorless, start in {1,2,3,4,5,6,7,8} e.g., Right goes to {2,4,6,8} Solution? [Right,Vacuum,Left,Vacuum] Contingency –Nondeterministic: Vacuum may dirty a clean carpet –Partially observable: location, dirt at current location. –Percept: [L, Clean], i.e., start in #5 or #7 Solution? [Right, if dirt then Vacuum]

10. Formal Problem Definitions Key components: –Initial state: E.g. First location –Available actions: Successor function: reachable states –Goal test: Conditions for goal satisfaction –Path cost: Cost of sequence from initial state to reachable state Solution: Path from initial state to goal –Optimal if lowest cost

11. Selecting a state space Real world is absurdly complex  state space must be abstracted for problem solving (Abstract) state = set of real states (Abstract) action = complex combination of real actions –e.g., "Arad  Zerind" represents a complex set of possible routes, detours, rest stops, etc. For guaranteed realizability, any real state "in Arad“ must get to some real state "in Zerind" (Abstract) solution = –set of real paths that are solutions in the real world Each abstract action should be "easier" than the original problem

12. Why Search? Not just city route search –Many AI problems can be posed as search Planning: –Vertices: World states; Edges: Actions Game-playing: –Vertices: Board configurations; Edges: Moves Speech Recognition: –Vertices: Phonemes; Edges: Phone transitions

13. Vacuum world state space graph states? actions? goal test? path cost?

14. Vacuum world state space graph states? integer dirt and robot location actions? Left, Right, Vacuum goal test? no dirt at all locations path cost? 1 per action

15. Example: The 8-puzzle states? actions? goal test? path cost?

16. Example: The 8-puzzle states? locations of tiles actions? move blank left, right, up, down goal test? = goal state (given) path cost? 1 per move [Note: optimal solution of n-Puzzle family is NP-hard]

17. Example: robotic assembly states?: real-valued coordinates of robot joint angles, parts of the object to be assembled actions?: continuous motions of robot joints goal test?: complete assembly path cost?: time to execute

18. Basic Search Algorithm Form a 1-element queue of 0 cost=root node Until first path in queue ends at goal or no paths –Remove 1st path from queue; extend path one step –Reject all paths with loops –Add new paths to queue If goal found=>success; else, failure Paths to extend Order of paths added Position new paths added

20. Basic Search Problem Vertices: Cities; Edges: Steps to next, distance Find route from S(tart) to G(oal) SDAEBCFG 3 4 5 4 24 3 4 5

21. Formal Statement Initial State: in(S) Successor function: –Go to all neighboring nodes Goal state: in(G) Path cost: –Sum of edge costs

25. Blind Search Need SOME route from S to G –Assume no information known –Depth-first search, breadth-first search, uniform-cost search, iterative deepening search Convert search problem to search tree –Root=Zero length path at Start –Node=Path: label by terminal node Child one-step extension of parent path

26. Search Tree SADBDCEDFGEFBGCEABDEFGBFACG

27. Breadth-first Search Explore all paths to a given depth SADBDAECEEBBFDFBFDEACG

28. Breadth-first Search Algorithm Form a 1-element queue of 0 cost=root node Until first path in queue ends at goal or no paths –Remove 1st path from queue; extend path one step –Reject all paths with loops –Add new paths to BACK of queue If goal found=>success; else, failure

29. Analyzing Search Algorithms Criteria: –Completeness: Finds a solution if one exists –Optimal: Find the best (least cost) solution –Time complexity: Order of growth of running time –Space complexity: Order of growth of space needs BFS: –Complete: yes; Optimal: only if # steps= cost –Time complexity: O(b^d+1); Space: O(b^d+1)

30. Uniform-cost Search BFS: –Extends path with fewest steps UCS: –Extends path with least cost Analysis: –Complete?: Yes; Optimal?: Yes –Time: O(b^(C*/e)); Space: O(b^(C*/e))

31. Uniform-cost Search Algorithm Form a 1-element queue of 0 cost=root node Until first path in queue ends at goal or no paths –Remove 1st path from queue; extend path one step –Reject all paths with loops –Add new paths to queue –Sort paths in order of increasing length If goal found=>success; else, failure

32. Depth-first Search Pick a child of each node visited, go forward –Ignore alternatives until exhaust path w/o goal SABCEDFG

33. Depth-first Search Algorithm Form a 1-element queue of 0 cost=root node Until first path in queue ends at goal or no paths –Remove 1st path from queue; extend path one step –Reject all paths with loops –Add new paths to FRONT of queue If goal found=>success; else, failure

34. Search Issues Breadth-first search: –Good if many (effectively) infinite paths, b<< –Bad if many end at same short depth, b>> Uniform-cost search: –Tries paths with many short steps first –Identical to BFS if all steps same cost Depth-first search: –Good if: most partial=>complete, not too long –Bad if many (effectively) infinite paths

35. Iterative Deepening Problem: –DFS good space behavior Could go down blind path, or sub-optimal Solution: –Search at progressively greater depths: 1,2,3,4,5…..

36. Progressive Deepening Question: Aren’t we wasting a lot of work? –E.g. cost of intermediate depths Answer: (surprisingly) No! –Most nodes at fringe –Fringe (depth d): Cost = b^d –Preceding depths: b^0 + b^1+…b^(d-1) (b^d - 1)/(b -1) –Ratio of last ply cost/all preceding ~ b - 1 –For large branching factors, prior work small relative to maximum depth

37. Heuristic Search A little knowledge is a powerful thing –Order choices to explore better options first –More knowledge => less search –Better search alg?? Better search space Measure of remaining cost to goal-heuristic –E.g. actual distance => straight-line distance S D A E BC F G 4.0 6.7 10.4 11.0 8.9 6.9 3.0

38. Hill-climbing Search Select child to expand that is closest to goal S A 10.4 D 8.9 A 10.4 E 6.9 B 6.7 F 3.0 G

39. Hill-climbing Search Algorithm Form a 1-element queue of 0 cost=root node Until first path in queue ends at goal or no paths –Remove 1st path from queue; extend path one step –Reject all paths with loops –Sort new paths by estimated distance to goal –Add new paths to FRONT of queue If goal found=>success; else, failure

40. Beam Search Breadth-first search of fixed width - top w –Guarantees limited branching factor, E.g. w=2 S A D BD AE 10.48.9 6.7 8.910.4 6.9 CE BF 4.0 6.9 6.73 ACG

41. Beam Search Algorithm –Form a 1-element queue of 0 cost=root node –Until first path in queue ends at goal or no paths Extend all paths one step Reject all paths with loops Sort all paths in queue by estimated distance to goal –Put top w in queue –If goal found=>success; else, failure

42. Best-first Search Expand best open node ANYWHERE in tree –Form a 1-element queue of 0 cost=root node –Until first path in queue ends at goal or no paths Remove 1st path from queue; extend path one step Reject all paths with loops Put in queue Sort all paths by estimated distance to goal –If goal found=>success; else, failure

43. Heuristic Search Issues Parameter-oriented hill climbing –Make one step adjustments to all parameters E.g. tuning brightness, contrast, r, g, b on TV –Test effect on performance measure Problems: –Foothill problem: aka local maximum All one-step changes - worse!, but not global max –Plateau problem: one-step changes, no FOM + –Ridge problem: all one-steps down, but not even local max Solution (local max): Randomize!!

44. Summary Blind search: –Find some path to goal –Depth-first, Breadth-first –Iterative Deepening Analyzing search algorithms –Completeness, Optimality, Time, Space Next time: –A little knowledge is a useful thing...

45. Search Costs