Artificial Intelligence Expert Systems Prolog

1
Artificial Intelligence / Expert SystemsProlog

• Justin Gaudry
• June 19, 2007

2
Logic Programming

• Horn clauses
• Formalized notation to express logical statements
• Negative premises (potentially) and positive
  conclusion (potentially) ORed together
• At most one positive literal (component)

3
Horn Clauses

• 3 forms
• No positive premise (conclusions to be
  discovered)
• p ? q ? r
• No negative conclusions (fact)
• C1
• At most one positive conclusion and all negative
  premises (rule)
• C1 ? p ? q ? r

4
Horn Clauses

• Modify
• C1 ? p ? q ? r
• p ? q ? r ? C1
• (p ? q ? r) ? C1
• (p ? q ? r) ? C1
• ? together premises and imply the conclusion

5
Prolog

• Horn clauses directly translate into Prolog
• Goals to be achieved queries
• Facts
• Rules

6
Rules

• Rules such as
• (p ? q ? r) ? C1
• translate to
• C1 - p, q, r.
• Facts such as
• C1
• translate to
• C1.

7
Backward Chaining

• Starting with conclusion, prove premises one by
  one until no goals left
• When a premise fails, backtracks and another is
  tried until no others can be found to try at
  which point the argument at that point fails and
  further backtracking higher up may be tried

8
Cut

• !
• Dont backtrack, fail the parent predicate.
• Use when only one direction is desired from a
  certain point on
• Allows for nonchronological backtracking

9

• holiday(friday, may1).
• weather(friday, fair).
• weather(saturday, fair).
• weather(sunday, fair).
• weekend(saturday).
• weekend(sunday).
• We go for picnics on good weekends and May 1st
• picnic(Day) - weather(Day,fair), weekend(Day).
• picnic(Day) - holiday(Day,may1).

10
Examples

• Do not look at the solution to Project 2
• (site not provided)
• http//www.unf.edu/jgaudry/Courses/acadinteg.html
  
• http//www.cs.nuim.ie/jpower/Courses/PROLOG/
• http//www.sics.se/isl/sicstuswww/site/links.html