Report

CS B551: ELEMENTS OF ARTIFICIAL INTELLIGENCE 1 Instructor: Kris Hauser http://cs.indiana.edu/~hauserk RECAP http://www.cs.indiana.edu/classes/b551 Brief history and philosophy of AI What is intelligence? Can a machine act/think intelligently? Turing machine, Chinese room 2 AGENDA Problem Solving using Search Search Algorithms 3 EXAMPLE: 8-PUZZLE 8 2 3 4 5 1 1 2 3 7 4 5 6 6 7 8 Initial state Goal state State: Any arrangement of 8 numbered tiles and an empty tile on a 3x3 board 4 SUCCESSOR FUNCTION: 8-PUZZLE 2 3 4 5 1 6 8 2 7 8 6 3 8 2 3 4 7 3 4 5 1 6 5 1 7 SUCC(state) subset of states 8 The successor function is knowledge about the 8-puzzle game, but it does not tell us which outcome to use, nor to which state of the board to apply it. 5 2 7 4 1 6 5 Across history, puzzles and games requiring the exploration of alternatives have been considered a challenge for human intelligence: Chess originated in Persia and India about 4000 years ago Checkers appear in 3600-year-old Egyptian paintings Go originated in China over 3000 years ago So, it’s not surprising that AI uses games 6 to design and test algorithms EXPLORING ALTERNATIVES Problems that seem to require intelligence usually require exploring multiple alternatives Search: a systematic way of exploring alternatives 7 8-QUEENS PROBLEM State repr. 1 Any non-conflicting placement of 0-8 queens State repr. 2 Any placement of 8 queens 8 DEFINING A SEARCH PROBLEM S State space S Successor function: x S SUCC(x) 2S Initial state s0 Goal test: xS GOAL?(x) =T or F Arc cost 9 STATE GRAPH Each state is represented by a distinct node An arc (or edge) connects a node s to a node s’ if s’ SUCC(s) The state graph may contain more than one connected component 10 SOLUTION TO THE SEARCH PROBLEM A solution is a path connecting the initial node to a goal node (any one) The cost of a path is the sum of the arc costs along this path An optimal solution is a solution path of minimum cost There might be no solution ! G I 11 PATHLESS PROBLEMS Sometimes the path doesn’t matter A solution is any goal node Arcs represent potential state transformations E.g. 8-queens, Simplex for LPs, Map coloring G I 12 REPRESENTATION 1 State: any placement of 0-8 queens Initial state: 0 queens Successor function: Goal test: Place queen in empty square Non-conflicting placement of 8 queens # of states ~ 64x63x…x57 ~ 3x1014 13 REPRESENTATION 2 State: any placement of nonconflicting 0-8 queens in columns starting from left Initial state: 0 queens Successor function: Goal test: A queen placed in leftmost empty column such that it causes no conflicts Any state with 8 queens # of states = 2057 14 PATH PLANNING What is the state space? 15 FORMULATION #1 Cost of one horizontal/vertical step = 1 Cost of one diagonal step = 2 16 OPTIMAL SOLUTION This path is the shortest in the discretized state space, but not in the original continuous space 17 FORMULATION #2 Cost of one step: length of segment 18 FORMULATION #2 Visibility graph Cost of one step: length of segment 19 SOLUTION PATH The shortest path in this state space is also the shortest in the original continuous space 20 WHAT IS A STATE? A state does: Represent all information meaningful to the problem at a given “instant in time” – past, present, or future Exist in an abstract, mathematical sense A state DOES NOT: Necessarily exist in the computer’s memory Tell the computer how it arrived at the state Tell the computer how to choose the next state Need to be a unique representation 21 WHAT IS A STATE SPACE? An abstract mathematical object Membership should be trivially testable E.g., the set of all permutations of (1,…,8,empty) E.g., S = { s | s is reachable from the start state through transformations of the successor function } is not easily testable Arcs should be easily generated Again: the state space does NOT contain information about which arc to take (or not to take) in a given state 22 5-MINUTE QUIZ Formulate 2x2 Tic-Tac-Toe, where you play both X and O’s actions, as a search problem. The goal state is any state with two in a line. Assume O goes first. Draw entire state graph. For compactness’s sake, eliminate symmetrical states Indicate initial and goal states on this graph Suppose one side is allowed to pass. How does the state graph change? Do you need to change anything to the problem definition? 23 EXAMPLE: 8-PUZZLE 8 2 3 4 5 1 1 2 3 7 4 5 6 6 7 8 Initial state Goal state State: Any arrangement of 8 numbered tiles and an empty tile on a 3x3 board 24 15-PUZZLE Introduced (?) in 1878 by Sam Loyd, who dubbed himself “America’s greatest puzzleexpert” 25 15-PUZZLE Sam Loyd offered $1,000 of his own money to the first person who would solve the following problem: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 ? 1 2 3 4 5 6 7 8 9 10 11 12 13 15 14 26 But no one ever won the prize !! 27 HOW BIG IS THE STATE SPACE OF THE (N21)-PUZZLE? 8-puzzle ?? states 28 HOW BIG IS THE STATE SPACE OF THE (N21)-PUZZLE? 8-puzzle 9! = 362,880 states 15-puzzle 16! ~ 2.09 x 1013 states 24-puzzle 25! ~ 1025 states But only half of these states are reachable from any given state (but you may not know that in advance) 29 PERMUTATION INVERSIONS Wlg, let the goal be: 1 5 9 13 2 6 10 14 3 4 7 8 11 12 15 A tile j appears after a tile i if either j appears on the same row as i to the right of i, or on another row below the row of i. For every i = 1, 2, ..., 15, let ni be the number of tiles j < i that appear after tile i (permutation inversions) N = n2 + n3 + + n15 + row number of empty tile 1 3 4 5 10 7 8 9 2 6 11 12 13 14 15 n2 = 0 n5 = 0 n8 = 1 n11 = 0 n14 = 0 n3 = 0 n6 = 0 n9 = 1 n12 = 0 n15 = 0 n4 = 0 n7 = 1 n10 = 4 n13 = 0 N=7+4 30 Proposition: (N mod 2) is invariant under any legal move of the empty tile Proof: Any horizontal move of the empty tile leaves N unchanged A vertical move of the empty tile changes N by an even increment ( 1 1 1 1) s= 1 2 5 6 3 4 1 2 7 5 6 11 7 9 10 11 8 13 14 15 12 s’ = 9 10 3 4 8 13 14 15 12 N(s’) = N(s) + 3 + 1 31 Proposition: (N mod 2) is invariant under any legal move of the empty tile For a goal state g to be reachable from a state s, a necessary condition is that N(g) and N(s) have the same parity It can be shown that this is also a sufficient condition The state graph consists of two connected components of equal size 32 SEARCHING THE STATE SPACE It is often not feasible (or too expensive) to build a complete representation of the state graph 33 8-, 15-, 24-PUZZLES 8-puzzle 362,880 states 0.036 sec 15-puzzle 2.09 x 1013 states ~ 55 hours 24-puzzle 1025 states > 109 years 100 millions states/sec 34 INTRACTABILITY Constructing the full state graph is intractable for most interesting problems n-puzzle: (n+1)! states k-queens: kk states Tractability of search hinges on the ability to explore only a tiny portion of the state graph! 35 SEARCHING 36 SEARCHING THE STATE SPACE Search tree 37 SEARCHING THE STATE SPACE Search tree 38 SEARCHING THE STATE SPACE Search tree 39 SEARCHING THE STATE SPACE Search tree 40 SEARCHING THE STATE SPACE Search tree 41 SEARCHING THE STATE SPACE Search tree 42 SEARCH NODES AND STATES 8 2 3 4 7 5 1 6 8 2 7 3 4 5 1 If states are allowed to be revisited, the search tree may be infinite even when the state space is finite 6 8 2 8 2 8 3 4 7 3 4 7 3 5 1 6 5 1 6 5 4 1 2 8 2 7 3 4 6 5 1 43 6 7 DATA STRUCTURE OF A NODE 8 2 3 4 7 5 1 6 STATE PARENT-NODE BOOKKEEPING CHILDREN ... Action Right Depth 5 Path-Cost 5 Expanded yes Depth of a node N = length of path from root to N (depth of the root = 0) 44 8 NODE EXPANSION The expansion of a node N of the search tree consists of: node generation node expansion 8 2 3 4 7 5 1 6 N Evaluating the successor function on STATE(N) Generating a child of N for each state returned by the function 2 8 4 2 7 3 4 7 3 5 1 6 5 1 6 8 2 3 4 7 5 1 45 6 FRINGE OF SEARCH TREE The fringe is the set of all search nodes that haven’t been expanded yet 8 2 3 4 7 5 1 6 8 2 7 3 4 5 1 6 8 2 3 4 7 5 1 6 8 2 3 4 7 5 1 6 8 2 7 3 4 5 1 6 8 2 3 4 7 5 1 6 Is it identical to the set of leaves? 47 SEARCH STRATEGY The fringe is the set of all search nodes that haven’t been expanded yet The fringe is implemented as a priority queue FRINGE INSERT(node,FRINGE) REMOVE(FRINGE) The ordering of the nodes in FRINGE defines the search strategy 48 SEARCH ALGORITHM #1 SEARCH#1 1. If GOAL?(initial-state) then return initial-state 2. INSERT(initial-node,FRINGE) 3. Repeat: 4. If empty(FRINGE) then return failure 5. N REMOVE(FRINGE) Expansion of N 6. s STATE(N) 7. For every state s’ in SUCCESSORS(s) 8. Create a new node N’ as a child of N 9. If GOAL?(s’) then return path or goal state 10. INSERT(N’,FRINGE) 49 PERFORMANCE MEASURES Completeness A search algorithm is complete if it finds a solution whenever one exists [What about the case when no solution exists?] Optimality A search algorithm is optimal if it returns a minimum-cost path whenever a solution exists Complexity It measures the time and amount of memory required by the algorithm 50 TOPICS OF NEXT 3-4 CLASSES Blind (uninformed) Search Heuristic (informed) Search Little or no knowledge about how to search How to use extra knowledge about the problem Local Search With knowledge about goal distribution 51 RECAP General problem solving framework State space Successor function Goal test => State graph Search is a methodical way of exploring alternatives 52 HOMEWORK Register! Readings: R&N Ch. 3.4-3.5 HW1 On OnCourse Writing and programming Due date: 9/6 53