CSCE 580 Artificial Intelligence Problem Spaces and Search

1
CSCE 580Artificial IntelligenceProblem Spaces
and Search

• Fall 2008
• Jingsong Wang
• wang82_at_engr.sc.edu

2
Abstraction and Generality in Java

• The Interaction between AI theory and program
  architecture
• Implementation of search strategies
• Generality and reuse
• Abstraction mechanisms and design patterns

3
Search Algorithms

• Uninformed algorithms
• Depth-first search (DFS)
• Breadth-first search (BFS)
• Informed algorithms
• Best-first search

4
Search Algorithms

• Model the structure of a problem graphically as a
  statespace
• Elements defining a state-space
• A formal representation of possible states of a
  problem solution
• Operators for generating new states from a given
  state
• Some way of recognizing a goal state
• A starting state of the problem, represented as
  the root of the graph

5
(No Transcript)
6
(No Transcript)
7
Search Algorithms

• DFS stack
• BFS queue
• Best-first search priority queue
• Heuristic rank
• sorted

8
Abstracting Problem States

• The separation of representation and control
• States and search engines
• Goal to define an abstract representation of
  problems states that supports the general search
  algorithm and can be easily specialized through
  the mechanism of class inheritance

9
Abstracting Problem States
10
Abstracting Problem States

• Abstract class

11
Abstracting Problem States

• Interface

12
Abstracting Problem States

• Javas generic
• Collection-TypeltElement-Typegt

13
(No Transcript)
14
(No Transcript)
15
Traversing the Problem Space

• Implementation of search itself
• Define the list of states and the mechanisms for
  moving through them in search

16
(No Transcript)
17
(No Transcript)
18
(No Transcript)
19
(No Transcript)
20
Putting the Framework to Use

• See the code in Eclipse