Presentation is loading. Please wait.

Presentation is loading. Please wait.

Search Algorithms in Intelligent Agents Presented by H. Joshi and J. McAdams {hmjoshi, CPSC7321 – Distributed OS Final Presentation.

Similar presentations


Presentation on theme: "Search Algorithms in Intelligent Agents Presented by H. Joshi and J. McAdams {hmjoshi, CPSC7321 – Distributed OS Final Presentation."— Presentation transcript:

1 Search Algorithms in Intelligent Agents Presented by H. Joshi and J. McAdams {hmjoshi, jamcadams} @ualr.edu CPSC7321 – Distributed OS Final Presentation © 2003 UALR

2 Outline  Introduction  Search Algorithms Traditional Heuristic Asynchronous Algorithm Comparison  Semantic Search  Conclusion  Demo

3 Introduction  Current Algorithms Text Based File Name Based Frame Based – NLP  Semantic Search

4 Traditional Search Algorithms  Breadth-First  Depth-First  Depth-First, Iterative Deepening  Bidirectional

5 Breadth-First

6 Depth-First

7 Depth-First, Iterative Deepening

8 Heuristic Search Algorithms  Generate and Test  Hill-Climbing  Best-First  A*  Iterative Deepening A*  Beam  Means-Ends  Constraint

9 Generate and Test

10 Hill-Climbing

11 Best-First

12 A*

13 Asynchronous Search Algorithms  Asynchronous Dynamic Programming  Learning Real-Time A*  Real-Time Multi-agent

14 Asynchronous Dynamic Programming

15 Learning Real-Time A*

16 Features For Comparison  Latency of the network affecting search process  Searching for semantically correct information  Relevance of the information  Order of complexity of an algorithm  Suitable for Text, graphics, audio, video (multimedia) contents  Does the algorithm consider aging of information and outdated/obsolete information  Distributiveness of an algorithm

17 Algorithm Comparison AlgorithmTimeMemoryCompleteOptimal Breadth-FirstO(b d ) Yes Depth-FirstO(b d )O(d)No Depth-First Iterative-DeepeningO(b d )O(d)Yes BidirectionalO(b d/2 ) Yes Hill-ClimbingO(b d )O(1) – O(b d )No Best-FirstO(b d ) YesNo A*O(b d ) Yes Iterative-Deepening-A*O(b d )O(d)Yes Beam SearchO(n d ) No Means-EndsO(b d ) No Generate and TestO(((bd!)/((bd-d)!))/2)O(d)YesNo Asynchronous DynamicO(b d )O(1)Yes Learning Real-Time A*O(b d ) Yes Real-Time A*O(b d ) YesNo

18 Semantic Intelligent Search Agent

19 Conclusion  Having the right tool for the job  Meaningful searching

20 Demo and Questions


Download ppt "Search Algorithms in Intelligent Agents Presented by H. Joshi and J. McAdams {hmjoshi, CPSC7321 – Distributed OS Final Presentation."

Similar presentations


Ads by Google