1 / 49

ISC 4322/6300 – GAM 4322 Artificial Intelligence

ISC 4322/6300 – GAM 4322 Artificial Intelligence. Lecture 3 Informed Search and Exploration Instructor: Alireza Tavakkoli September 10, 2009. University Of Houston- Victoria Computer Science Department. Uninformed Search. Problem Solving Agents A sequence of actions Desirable state

ivana
Télécharger la présentation

ISC 4322/6300 – GAM 4322 Artificial Intelligence

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. ISC 4322/6300 – GAM 4322Artificial Intelligence Lecture 3 Informed Search and Exploration Instructor: Alireza Tavakkoli September 10, 2009 University Of Houston- Victoria Computer Science Department

  2. Uninformed Search • Problem Solving Agents • A sequence of actions • Desirable state • Definition • Problem • Solution • Approach • Uninformed search mechanisms • Analysis of algorithms concepts • Asymptotic complexity • NP-completeness

  3. Problem Solving • What is a goal? • A set of world states • What actions to take? • Level of detail • Problem formulation • In our example • Driving to a particular town • States are being in a particular town • Goal • Driving to Bucharest

  4. Searching for Solutions • Search tree • Search nodes • Root • Initial state search node • Expansion of states • Applying the successor function Search Strategy

  5. Comparing Search Strategies

  6. Summary of Algorithms

  7. Material • Chapter 4 Section 1 – 3 • Exclude memory-bounded heuristic search • Outline • Informed search strategies • Greedy best-first search • A* search • Heuristics • Local search algorithms • Hill-climbing search • Simulated annealing search • Local beam search • Genetic algorithms

  8. Informed Search 282 538 • So far the techniques made use only of information that could be stored on a graph, i.e., the successor function • Often, however, we have available problem-specific knowledge that can help. • When additional information is used... we talk about “Informed Search”. • How can we take advantage of such additional information?

  9. Informed Search Strategies • Informed search • Has more knowledge beyond the problem definition • Best-First Search • Expand the seemingly best node • Cost to the goal is minimum • Evaluation function f(n) • Data structure • Priority queue  a fringe in ascending order of f-values • Heuristic function h(n) • If n is the goal then h(n)= 0 • Variants • Greedy Best-First Search • A* Algorithm

  10. Romania w/ SLD in Km

  11. Greedy Best-First Search • Greedy best-first search expands the node that appears to be closest to goal • Evaluation function f(n) = h(n) (heuristic) • = estimate of cost from n to goal • e.g., hSLD(n) = straight-line distance from n to Bucharest • The heuristic function can not be computed from the problem definition • The heuristic function is relevant to the problem goal • Straight line distance is likely related to the road distance

  12. Greedy Best-First Search Example

  13. Greedy Best-First Search Example

  14. Greedy Best-First Search Example

  15. Greedy Best-First Search Example

  16. Properties of GBFS • Complete? • No – can get stuck in loops • e.g., Iasi  Neamt  Iasi  Neamt  • Time? • O(bm) • but a good heuristic can give dramatic improvement • Space? • O(bm) • keeps all nodes in memory • Optimal? • No

  17. A* Search • Idea: avoid expanding paths that are already expensive • Evaluation function f(n) = g(n) + h(n) • g(n) = cost so far to reach n • h(n) = estimated cost from n to goal • f(n) = estimated total cost of path through n to goal

  18. A* search example

  19. A* search example

  20. A* search example

  21. A* search example

  22. A* search example

  23. A* search example

  24. Admissible Heuristics • A heuristic h(n) is admissible if for every node n, h(n) ≤ h*(n), where h*(n) is the true cost to reach the goal state from n. • An admissible heuristic never overestimates the cost to reach the goal, i.e., it is optimistic • Example: hSLD(n) (never overestimates the actual road distance) • Optimality • Theorem: If h(n) is admissible, A* using TREE-SEARCH is optimal • Tree search allows for repeated states

  25. Optimality of A* (proof) • Suppose some suboptimal goal G2has been generated and is in the fringe. Let n be an unexpanded node in the fringe such that n is on a shortest path to an optimal goal G. • f(G2) = g(G2) since h(G2) = 0 • g(G2) > g(G) since G2 is suboptimal • f(G) = g(G) since h(G) = 0 • f(G2) > f(G) from above

  26. Optimality of A* (proof) • Suppose some suboptimal goal G2has been generated and is in the fringe. Let n be an unexpanded node in the fringe such that n is on a shortest path to an optimal goal G. • f(G2) > f(G) from above • h(n) ≤ h*(n) since h is admissible • g(n) + h(n) ≤ g(n) + h*(n) • f(n) ≤ f(G) < f(G2) Hence f(G2) > f(n), and A* will never select G2 for expansion

  27. A* Using Graph Search • Case 2: GRAPH-SEARCH • (avoid repeated states) • Admissibility is not sufficient. • 2 solutions to retain optimality: • once you found a repeated state, discard the more expensive path (extra book-keeping, expensive) • make sure the first path followed to a repeated state is the optimal path • Consistency • If h(n) is consistent • then A* is optimal • Admissible = Monotonic • Satisfies the triangular inequality

  28. Consistent heuristics • A heuristic is consistent if for every node n, every successor n' of n generated by any action a, • h(n) ≤ c(n,a,n') + h(n') • If h is consistent, we have 1- f(n') = g(n') + h(n')  2- f(n') = g(n) + c(n,a,n') + h(n'), 3- h(n) ≤ c(n,a,n') + h(n')  4- g(n)+ c(n,a,n') + h(n') ≥ g(n) + h(n)  From 2 & 4- f(n') ≥ f(n) • i.e., f(n) is non-decreasing along any path. • Theorem: If h(n) is consistent, A* using GRAPH-SEARCH is optimal • Graph search does not allow for repeated states

  29. Optimality of A* • A* expands nodes in order of increasing f value • Gradually adds "f-contours" of nodes • Contour i has all nodes with f=fi, where fi < fi+1

  30. Properties of A* • Complete? • Yes (unless there are infinitely many nodes with f ≤ f(G) ) • Time? • Exponential • Space? • Keeps all nodes in memory • Optimal? • Yes

  31. Admissible Heuristics E.g., for the 8-puzzle: • h1(n) = number of misplaced tiles • h2(n) = total Manhattan distance (i.e., no. of squares from desired location of each tile) • h1(S) = ? • h2(S) = ?

  32. Admissible Heuristics E.g., for the 8-puzzle: • h1(n) = number of misplaced tiles • h2(n) = total Manhattan distance (i.e., no. of squares from desired location of each tile) • h1(S) = ? 8 • h2(S) = ? 3+1+2+2+2+3+3+2 = 18 • Both heuristics are admissible.

  33. Dominance • If h2(n) ≥ h1(n) for all n (both admissible) • then h2dominatesh1 • h2is better for search • Effective branching factor b* • Good heuristics lead to b*1 • Typical search costs (average number of nodes expanded): • Check table 4.8 • d=12 IDS = 3,644,035 nodes b*= 2.78 A*(h1) = 227 nodes b*= 1.42 A*(h2) = 73 nodes b*= 1.24 • d=24 IDS = too many nodes b*= NA A*(h1) = 39,135 nodes b*= 1.48 A*(h2) = 1,641 nodes b*= 1.26

  34. Relaxed Problems • A problem with fewer restrictions on the actions is called a relaxed problem • The cost of an optimal solution to a relaxed problem is an admissible heuristic for the original problem • “The optimal solution to the original problem is, by definition, also a solution to the relaxed problem and therefore, must be at least as expensive as the optimal solution to the relaxed problem. • 8-Puzzle relaxed problems • If the rules of the 8-puzzle are relaxed so that a tile can move anywhere, then h1(n) gives the shortest solution • If the rules are relaxed so that a tile can move to any adjacent square, then h2(n) gives the shortest solution

  35. Local Search Algorithms • In many optimization problems, the path to the goal is irrelevant • the goal state itself is the solution • State space = set of "complete" configurations • Find configuration satisfying constraints • e.g., n-queens • In such cases, we can use local search algorithms • keep a single "current" state, try to improve it • Advantages: • Very little memory requirements • Often find reasonable solutions in large spaces • Suitable for optimization problems

  36. Example: n-queens • Put n queens on an n × n board with no two queens on the same row, column, or diagonal

  37. Hill-climbing search • State space landscape

  38. Hill-climbing search • "Like climbing Everest in thick fog with amnesia"

  39. Hill-climbing search • Problem: depending on initial state, can get stuck in local maxima

  40. Hill-climbing search: 8-queens problem • How many successor each state has? • 56 • h = number of pairs of queens that are attacking each other, either directly or indirectly • h = 17 for the above state

  41. Hill-climbing search: 8-queens problem • Greed algorithms work well • They make quick progress • However • They get stuck in local maxima, minima • They don’t look ahead! Almost a perfect solution with h=1 Took only 5 steps from the previous Slide!

  42. When does hill-climbing fail? • Local maxima/minima • Not found a global maxima, but the neighboring stares are even worse! • Ridges • All maxima are the same and everything else is worse! • Plateaux • All neighboring states have the same evaluation value!

  43. Simulated Annealing Search • Idea: escape local maxima by allowing some "bad" moves but gradually decrease their frequency

  44. Properties of Simulated Annealing • One can prove: If T decreases slowly enough, then simulated annealing search will find a global optimum with probability approaching 1 • Widely used in VLSI layout, airline scheduling, etc

  45. Local Beam Search • Keep track of k states rather than just one • Start with k randomly generated states • At each iteration, all the successors of all k states are generated • If any one is a goal state, stop; else select the k best successors from the complete list and repeat.

  46. Genetic Algorithms • A successor state is generated by combining two parent states • Start with k randomly generated states (population) • A state is represented as a string over a finite alphabet (often a string of 0s and 1s) • Evaluation function (fitness function). Higher values for better states. • Produce the next generation of states by selection, crossover, and mutation

  47. Genetic algorithms • Fitness function: number of non-attacking pairs of queens (min = 0, max = 8 × 7/2 = 28) • 24/(24+23+20+11) = 31% • 23/(24+23+20+11) = 29% etc

  48. Genetic algorithms

  49. Questions? • AI: A Modern Approach : • Chapter 4

More Related