1 / 13

Artificial Intelligence

Artificial Intelligence. Informed Search Methods. Chapter 4. Review. Last week, we talked about: Uninformed search strategies Can find solutions to problems by generating new states Testing them against the goal.

march
Télécharger la présentation

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. Artificial Intelligence Informed Search Methods Chapter 4

  2. Review Last week, we talked about: • Uninformed search strategies • Can find solutions to problems by generating new states • Testing them against the goal. • uses no information that may be available about the structureof the tree or availability of goals, or any other information that may help optimise the search. • Inefficient in most case This week: • Informed search strategies: • Uses problem-specific knowledge • makes use of available informationmaking the search more efficient.

  3. Outline of this Chapter • Best-first search • Greedy best-first search • A* search • Example of Heuristics

  4. Informed search • Usually more efficient than blind search • It is also called a Heuristic search • A heuristic is a rule or method that almost alwaysimproves the decision process. • Sometimes state space is too large to search thoroughly • it is possible to introduce a scoring function, or evaluation function, that estimates which nodes seem promising • focus on paths that consist of promising nodes

  5. Definition of a Heuristic Function • heuristic search- we explore the node that is most likely to be nearest to a goal state. To do this we use a heuristic function which tells us how close we are to a goal state , not how cheap the solution is so far. • A heuristic function is normally denoted h(n): • h(n) = estimated cost of the cheapest path from node n to a goal node.

  6. Best-First Search (BestFS) • Type of Informed/Heuristic search • Uses an evaluation function on all successor nodes • Node with the lowest evaluation is expanded first (evaluation measures distance to the goal). • A combination of depth first (DFS) and breadth first search (BFS). • keep track of all other nodes at same level (like breadth-first) & does not get trapped in dead ends • follow single, most promising path down through tree (like depth-first)  solution can be found without computing all nodes. • Special cases: • greedy best-first search • A* search

  7. Greedy best-first search • This is one of the simplest BestFS strategy • It tries to get to the goal as it can at each step, without worrying about whether this will be best in the long run. (see next slide for an example). • Greedy best-first search expands the node that appears to be closest to the goal state first lead to a solution quickly. • The cost of reaching the goal from a state can be estimated using h(n) but cannot be determined exactly. • e.g. Route-finding problem (RFD) • A good heuristic function of RFP is SLD( Straight Line Distance). That is, • hSLD(n) = straight-line distance from n to the goal

  8. Map of Romania with road distance in Km, & SLD to Bucharest We will see next the progress of GBFS using hSLD to find a path from A to B

  9. Straight Line Distance

  10. Greedy best-first search Example Solution found without expanding a node that is not on the solution path  search cost is minimal. Sol ={Arad, Sibiu, Fagaras, Bucharest} But the Best Sol ={Arad, Sibiu, Rimnicu Vilcea, Pitesti, Bucharest}

  11. Properties of greedy best-first search • It is not optimal- e.g. the path it founds via Sibiu & Fagaras to Bucharest is 32 km longer than path through Rimnicu Vilcea & Pitesti. • It is not complete- solution may never be found, It is possible to get stuck in an infinite loop (consider being in Iasi and trying to get to Fagaras) • The heuristic suggests that Neamt be expanded 1st because it is closest to Fagaras, but is a dead end. • Search will oscillate between Iasi  Neamt  Iasi  Neamt  • Time complexity is O(bm); where m is the depth of the search tree • SpaceO(bm)  keeps all nodes in memory

  12. A* search • Idea: avoid expanding paths that are already expensive. • Greedy search minimizes the estimated cost to the goal, h(n)  cuts the search cost but it is not optimal & complete. • UCS minimizes the costs of the path, g(n); it is optimal & complete but can be very inefficient. • To get the advantages of both strategies -- A* Search combinesthe 2 evaluation functions by summing them: f(n) = g(n) + h (n) g(n) = gives the path cost from the start node to node n h(n) = estimated cost from n to goal f(n) = estimated cost of the cheapest solution through n.

  13. A* search Example (Progress of A*tree search for Bucharest) The h values are the straight-line distances to Bucharest taken from Slide 8

More Related