Scribd
Upload
23 views

Logic Programming Tutorial 1: Prolog Warm-Up

This document provides instructions for exercises in a Prolog tutorial. It introduces concepts like unification, basic queries to retrieve facts about characters from The Simpsons, representing disjunction, adding symmetric rules, using negation and inequality to make relations irreflexive, and exploring goal ordering. Students are asked to define predicates to represent various relationships between characters and experiment with Prolog's behavior.

Uploaded by

Dr-Ashraf Heikal
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
23 views

Logic Programming Tutorial 1: Prolog Warm-Up

This document provides instructions for exercises in a Prolog tutorial. It introduces concepts like unification, basic queries to retrieve facts about characters from The Simpsons, representing disjunction, adding symmetric rules, using negation and inequality to make relations irreflexive, and exploring goal ordering. Students are asked to define predicates to represent various relationships between characters and experiment with Prolog's behavior.

Uploaded by

Dr-Ashraf Heikal
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

Logic Programming

Tutorial 1: Prolog Warm-up

You should be able to do all of the unstarred exercises before the tutorial session.
You should also be able to do exercises with a single star (*) but these may require
more thought. Exercises with two stars (**) are harder and may involve material
we haven’t covered; try them if you have time.

1. Unification
Without using a computer, find the most general unifiers of the following
unification problems (or determine that they are not unifiable).

(a) X = f(Y,Y), Y = f(Z,Z), Z = f(W,W)


(b) X = f(g(Y,Z), W), f(Y,W) = Z
(c) X = f(g(Y,Z), W), f(X,Z) = W
(d) X = f(Y), Y = f(Z), f(X,Y) = f(g(Z),W)

Compare your answers to Prolog’s behavior. Are there any surprises?


2. Basic queries
Start Sicstus and load the data in simpsons.pl, which defines some basic
facts about the charcters in the Simpsons television program.
Define predicates with the following behavior:
(a) person(X), which holds if X is a person. (It is safe to assume that all
people in this example are either male or female.)
(b) employer(X), which says that X is an employer (at least one person works
for X).
(c) parent(X,Y), which says that X is one of Y’s parents.
(d) grandparent(X,Y), which says that X is one of Y’s grandparents.
(e) sibling(X,Y), which holds if X and Y have a parent in common. (Tech-
nically, this should probably be called “half-sibling”.)
(f) classmate(X,Y), which says that X and Y are taught by the same person.
The predicates should have at most two rules and their bodies should be
conjunctions of atomic formulas.
Finally, write a goal that returns Bart’s sibling’s teacher’s employer.
3. Disjunction
The parent query in part (2) above involves disjunction (logical OR). In
Prolog, disjunction can be represented in two ways. First, we can define
multiple clauses that derive the same conclusion from different subgoals:

p or q :- p.
p or q :- q.

1
Alternatively, Prolog provides syntax for using disjunction within goals. Specif-
ically, if G1 and G2 are goals, we may write G1;G2 for the disjunction of the
two goals.

p or q :- p ; q.

Prolog will try to solve the first goal and if that fails (or if we backtrack) it
will then try the second goal. This syntax is helpful because sometimes we
can use this to avoid writing helper predicates to deal with disjunction.
Define predicate called parent disj/2 whose meaning is the same as parent/2
but uses disjunction. Compare its behavior (with and without tracing) to the
two-clause version.
4. Symmetry
The predicates friend/2 and neighbor/2 are not symmetric and they prob-
ably should be: if X is a friend or neighbor of Y then Y should be a friend or
neighbor of X as well.
One way to do this is simply to add rules such as

friend(X,Y) :- friend(Y,X).
neighbor(X,Y) :- neighbor(Y,X).

Add these rules, and write a query that searches for the catchphrases of Mil-
house’s friend’s parent’s neighbor. Try to find all solutions. Is there a potential
problem with this approach? How might we avoid this problem?
5. (*) Goal ordering
Define predicates aunt/2 and uncle/2 in terms of male, female, parent and
sibling. Observe the behavior of aunt(X,bart), with or without tracing.
Experiment with different goal orderings. Which goal ordering seems to be
the most efficient for the above query?
6. (**) Negation and inequality
The classmate and sibling predicates we defined above may give undesired
answers. Try solving classmate(X,X) or sibling(X,X) to see examples of
this problem.
To avoid this problem, we want to eliminate the pairs (X,X) from the classmate
relation to make it irreflexive. More generally, we often want to use negation
to make exceptions to rules.

(a) Prolog has a built-in non-unifiability predicate, written t \= u, that


succeeds if t and u do not unify. Use \= to implement a version of
classmate that is irreflexive. Do the same for sibling.
What happens now in response to queries ?- classmate(X,X) or ?-
classmate(X,Y)?
(b) Prolog also has a negation-as-failure primitive, written \+(G) that tries
to solve a goal G, and succeeds if G fails. Use negation-as-failure to define
a predicate people with no catchphrase/1 that holds of all people with
no catchphrase. Also, define a predicate adults with no catchphrase/1
that holds of all people who are not children and have no catchphrase.
This should be possible with a rule with only two goals (one negated).
Experiment with goal ordering — which ordering seems to have better
behavior?

You might also like