Back

25_Prolog_Advanced

Loading views...
← PreviousNext →

Prolog Advanced Control: Cut, Negation, and Built-ins

This section covers how to control Prolog's search behavior and use built-in predicates for input/output and database management.

1. The Cut (!) Operator

The Cut is a special predicate used to prevent backtracking. It "freezes" the choices made so far in the current rule.

Purpose:

  • Efficiency: Prune search branches that we know will not lead to a solution.
  • Mutually Exclusive Rules: Ensure that only one rule is applied (e.g., if-then-else behavior).

Example: Max of two numbers

max(X, Y, X) :- X >= Y, !.
max(X, Y, Y).
  • If X >= Y, the first rule matches, the cut is executed, and Prolog will not try the second rule even if the first rule eventually fails later.

2. Negation as Failure (not)

Prolog uses the Closed World Assumption: anything it cannot prove to be true is assumed to be false.

  • not(P): Succeeds if the goal P fails.
  • Limitation: It is not true logical negation. It only means "not provable".

3. Built-in Predicates

Prolog provides many pre-defined predicates for common tasks.

3.1 Input and Output (6.7)

  • read(X): Reads the next term from the input stream.
  • write(X): Outputs the term X to the output stream.
  • nl: Outputs a new line.
  • tab(N): Outputs N spaces.
  • see(File) / tell(File): Switches input/output to a file.
  • seen / told: Closes the current input/output file.

3.2 Database Manipulation

  • assert(P): Adds fact or rule P to the database at runtime.
  • retract(P): Removes P from the database.

4. Processing Files of Terms (6.7.2)

A common pattern is to read all terms from a file until the end.

Example:

process_file(File) :-
    see(File),
    repeat,
    read(Term),
    ( Term = end_of_file -> true ; process(Term), fail ),
    seen.
  • endoffile: A special atom returned by read when the file ends.

5. Summary of Prolog Programming Style

  1. Declarative: Focus on what is true.
  2. Recursive: Use recursion for loops.
  3. Symbolic: Handle data as symbols and structures.
  4. Controlled: Use the Cut (!) for performance and logic control.

4. Case Study: The Monkey-Banana Problem

The Monkey-Banana problem is a classic AI puzzle that demonstrates Prolog's power in planning and goal-oriented search.

Problem Scenario:

A monkey is in a room. Suspended from the ceiling are some bananas, out of the monkey's reach. Also in the room is a chair. The monkey can move around, push the chair, climb onto the chair, and grasp the bananas if it is on the chair under the bananas.

Prolog Implementation:

% state(MonkeyLocation, ChairLocation, HasBananas)

% 1. Grasping the bananas
move(state(middle, middle, hasnot), grasp, state(middle, middle, has)).

% 2. Climbing onto the chair
move(state(L, L, hasnot), climb, state(L, L, hasnot)).

% 3. Pushing the chair
move(state(L1, L1, hasnot), push(L1, L2), state(L2, L2, hasnot)).

% 4. Walking around
move(state(L1, L2, hasnot), walk(L1, L3), state(L3, L2, hasnot)).

% Recursive definition of a solution
canget(state(_, _, has)).
canget(State1) :-
    move(State1, Action, State2),
    canget(State2).
← PreviousNext →