chapter 1: introduction - ahmed sallam · prolog iv, binprolog, ciao prolog, prolog lpa, visual...
TRANSCRIPT
Ahmed Sallam Slides based on original lecture slides Dr. Temur Kutsia
LO
GIC
PR
OG
RA
MM
ING
Chapter 1: Introduction
"What" vs "How"
Declarative vs Procedural Programming
Procedural programming
• The programmer has to specify how to get the output for
the range of required inputs.
• The programmer must know the appropriate algorithm.
Declarative programming
• Requires a more descriptive style.
• The programmer must know what relationships hold
between various entities.
Introduction
2
16-Feb-17
Example: List Concatenation
In procedural style:
list procedure cat(list a,list b)
{
list t = list u = copylist(a);
while (t.tail != nil) t = t.tail;
t.tail = b;
return u;
}
In declarative style:
cat([], L, L).
cat([H|T], L, [H|Z]) :- cat(T, L, Z).
Introduction
3
16-Feb-17
Logic Programming
A declarative style programming paradigm.
Computation through logical deduction.
Uses the language of logic to express data and programs.
Most of current logic programming languages use first
order logic (FOL).
Prolog – the most popular logic programming language.
Introduction
4
16-Feb-17
Historical Facts
1970-ies:
• Bob Kowalski. "Predicate Logic as a Programming Language".
IFIP Congress, Stockholm
• Alain Colmerauer and his group. Interpreter of the first logic programming language Prolog.
Marseille
Introduction
5
16-Feb-17
Prolog
• Prolog is the main subject of this course
• Used in Artificial Intelligence, Natural Language
Processing, Automated Reasoning, XML Querying...
• Exists in many dialects (Sicstus Prolog, SWI Prolog,
Prolog IV, BinProlog, Ciao Prolog, Prolog LPA, Visual
Prolog, YAP Prolog, Strawberry Prolog)...
• (Almost) all the dialects agree on the “core" part (ISO
Standard for Prolog)
Introduction
6
16-Feb-17
Prolog in Industrial Applications
Clarissa:
• A fully voice-operated procedure browser.
• Developed at NASA.
• Used on the International Space Station.
• Enables astronauts to give full attention to the task while
they navigate through complex procedures using spoken
commands.
• Implemented in SICStus Prolog.
Introduction
7
16-Feb-17
Prolog in Industrial Applications
Some other solutions:
• FleetWatch – fully integrated operations control and schedules
planning solution. Used by 21 international airlines, among
them Comair (USA), Italian branch of Lauda Air, Malev
(Hungary), DHL Europe, Asiana (South Korea), Hainan (China),
Royal Jordanian, Kuwait Airways, Cimber Air (Denmark), etc.
• SCORE – a long term airport capacity management system for
coordinated airports. Successfully Used at IATA Scheduling
Conferences. Users in more than 20 countries.
• ARGOS – a Decision Support System (DSS) for enhancing
Crisis Management for incidents with Chemical, Biological,
Radiological, and Nuclear (CBRN) releases. Introduction
8
16-Feb-17
Contents
Basics of PROLOG
Facts
Questions
Variables
Conjunction
Rules
Introduction
9
16-Feb-17
PROLOG
Used to solve problems involving
• objects, and
• relationships between objects.
Introduction
10
16-Feb-17
Relationships
Example
John owns the book
• The relationship: ownership
• The objects: book, John
Directional:
• John owns the book
• Not: The book owns John
Introduction
11
16-Feb-17
PROLOG
Program can be thought of as a storehouse of facts and rules.
Conversational Language: The user can ask questions about
the set of facts and rules in the PROLOG program.
Introduction
12
16-Feb-17
PROLOG
Sisters Example:
A rule defining sisters and the facts about the people involved.
The user would ask:
Are these two people sisters?
The system would answer
yes (true) or no (false)
Introduction
13
16-Feb-17
Programming in PROLOG
Declaring Facts about objects and their relationships.
Defining Rules about objects and their relationships.
Asking Questions about objects and their relationships.
Introduction
14
16-Feb-17
Contents
Basics of PROLOG
Facts
Questions
Variables
Conjunction
Rules
Introduction
15
16-Feb-17
Parts of Fact
0
Introduction
16
16-Feb-17
Order of Objects
Introduction
17
16-Feb-17
Examples of Facts
Example
Gold is valuable. v a l u a b l e ( g o l d ) Jane is a female. f e m a l e ( j an e ) John owns some gold.
ow ns( john ,go ld ) John is the father of Mary.
f a t h e r ( j o h n , m a r y )
Are these expressions really facts? Is there anything missing?
Introduction
18
16-Feb-17
Interpretation of Names
The name refers to an object.
Semantic Meaning: Given by the programmer.
Syntactic Meaning: a set of characters, as PROLOG sees it.
Introduction
19
16-Feb-17
Interpretation of Names
Name refers to an object.
Name gold can refer to:
a particular lump of gold, or
the chemical element Gold having atomic number 79.
va luab le ( g o ld ) can mean:
that particular lump of gold, named gold, is valuable, or
the chemical element Gold, named gold, is valuable.
The programmer decides (in her usage) the meaning.
Introduction
20
16-Feb-17
Fact Terminology
Introduction
21
16-Feb-17
Database
Definition In PROLOG, database is a collection of facts.
PROLOG draws its knowledge from these facts.
The programmer is responsible for their accuracy.
Introduction
22
16-Feb-17
Contents
Basics of PROLOG
Facts
Questions
Variables
Conjunction
Rules
Introduction
23
16-Feb-17
Questions
The database contains the facts from which the questions are
answered.
A Question can look exactly like a fact:
owns(mary,book) .
The difference is in which mode one is in
Introduction
24
16-Feb-17
Questions
In the interactive question mode (indicated by the question
mark and dash): ? -
Question: ? - owns(mary,book).
Meanins: If mary is interpreted as a person called Mary, and book is
interpreted as some particular book, then
?- owns(mary,book).
means: Does Mary own the book?
Introduction
25
16-Feb-17
Database Search
Example
Facts in the database:
l i k e s ( j o e , f i s h ) .
l i k e s ( j o e , m a r y ) .
l i k es ( m ar y , book ) .
l i k e s ( j o h n , b o o k ) .
Questions: ?- l i k es ( j oe , m oney) . no
?- l i k e s ( j o e , m a r y ) . yes
?- k i n g ( j o h n , f r a n c e ) . n o
Introduction
26
16-Feb-17
Knowledge
The questions are always answered with respect to the
database.
Example
Facts in the database:
human(socrates).
h u m a n ( a r i s t o t l e ) .
a t hen ian ( soc r a tes ) .
Question:
Is Socrates Greek?
?- g reek (soc ra tes )
The answer with respect to this database is No.
Introduction
27
16-Feb-17
Questions
Up until now questions just reflect exactly the database.
Does Mary like the book?
?- l i k es ( m ar y , book ) . More Interesting Question:
What objects does Mary like?
Variables.
Introduction
28
16-Feb-17
Contents
Basics of PROLOG
Facts
Questions
Variables
Conjunction
Rules
Introduction
29
16-Feb-17
Variables
Tiresome to ask about every object:
l i k e s ( j o h n , t h i s ) .
l i k e s ( j o h n , t h a t ) .
Better to ask: What does John like?
or
Does John like X?
(i.e. use variables)
Introduction
30
16-Feb-17
Question With Variables
Does John like X?
? - l i k e s ( j o h n , X ) .
or
? - l i kes( john,Someth ingThatJohnL ikes ) .
X and SomethingThatJohnLikes are variables.
Variable begins with a capital letter.
Introduction
31
16-Feb-17
PROLOG Answer
Database:
l i k e s ( j o h n , f l o w e r s ) .
Question:
? - l i k e s ( j o h n , X ) .
PROLOG answers:
X=f lowers
Introduction
32
16-Feb-17
Many Answers
Database:
l i k e s ( j o h n , f l o w e r s ) .
l i k e s ( j o h n , m a r y ) .
l i k e s ( p a u l , m a r y ) .
Question:
? - l i k e s ( j o h n , X ) .
PROLOG answers:
X=f lowers
and the user acknowledges
X=mary
and the user acknowledges
no
Introduction
33
16-Feb-17
Placemarker
The first match is found: X=f lowers .
The user acknowledges.
From that place on the next match is found (the search continues).
From the place of the last instantiation no more match was found.
Thus answer: no.
Introduction
34
16-Feb-17
Contents
Basics of PROLOG
Facts
Questions
Variables
Conjunction
Rules
Introduction
35
16-Feb-17
Conjunctions
More Complicated Relationships:
Does Mary like John and does John like Mary?
Both Conditions must be fulfilled.
Introduction
36
16-Feb-17
Conjunctions
Database:
l i k e s ( m a r y , f o o d ) .
l i k e s ( m a r y , c o l a ) .
l i k e s ( j o h n , c o l a ) .
l i k e s ( j o h n , m a r y ) .
Comma means Conjunction:
? - l i k e s ( j o h n , m a r y ) , l i k e s ( m a r y , j o h n ) .
Answer: no
A match for l i k es ( j ohn , m a r y )
but none for l i k es ( m ar y , j ohn )
Introduction
37
16-Feb-17
Conjunctions with Variables
Is there anything that both mary and john like?
Find out what Mary likes and then see if John likes it.
? - l i k e s ( m a r y , X ) , l i k e s ( j o h n , X ) .
Introduction
38
16-Feb-17
Backtracking
Find match for the first goal.
Then see if matches the second.
If not, find another match for the first.
See if this matches the second.
…etc.
Introduction
39
16-Feb-17
Match First
Introduction
40
16-Feb-17
,
Cola).
,
Cola).
Match Second
Introduction
41
16-Feb-17
,
Cola). ,
Cola).
Backtrack
Introduction
42
16-Feb-17
,
Cola). ,
Cola).
Cola
Cola
Success
Introduction
43
16-Feb-17
,
Cola). ,
Cola).
Cola
Cola
Contents
Basics of PROLOG
Facts
Questions
Variables
Conjunction
Rules
Introduction
44
16-Feb-17
Rules
• How to express that John likes all people?
• Listing all people?
• likes(john, alfred).
• likes(john, bertrand).
• likes(john, charles).
• likes(john, david).
• etc.
• Not feasible. More compact way: Using rules.
John likes any object provided it is a person.
Introduction
45
16-Feb-17
Rule Examples
Rules state Dependence:
I use an umbrella if there is rain.
Rules Define:
X is a bird if X is an animal and X has feathers.
Introduction
46
16-Feb-17
Formulating Rules
John likes anyone who likes cola.
John likes any Something if it likes cola
John likes X if X likes cola
Introduction
47
16-Feb-17
Rule Syntax
Introduction
48
16-Feb-17
, Cola).
Variable Scope
The occurrences of X within a rule refer to the same object:
l i k e s ( j o h n , X ) : - l i k e s ( X , c o la ) , l i k e s ( X , f o o d ) .
l i k e s ( j o h n , mary):-
l i k e s ( m a r y , c o la ) , l i k e s ( m a r y , f o o d ) .
l i k e s ( j o h n , adam):- l i k es ( adam , c o l a ) , l i k es ( adam,
f o o d ) .
Introduction
49
16-Feb-17
Royal Parents
Example
The parents of X are Y and Z.
Y is the mother.
Z is the father.
Database:
m a l e ( a l b e r t ) .
male(edward).
f e m a l e ( a l i c e ) .
f e m a l e ( v i c t o r i a ) .
p a r e n t s ( e d w a r d , v i c t o r i a , a l b e r t ) .
p a r e n t s ( a l i c e , v i c t o r i a , a l b e r t ) . Introduction
50
16-Feb-17
Sisters
Example
X is a sister of Y if:
X is female,
X has parents M and F,
Y has parents M and F.
Rule:
s i s t e r ( X , Y ) : - f ema le (X ) ,
pa rents (X ,M,F ) ,
pa rents (Y ,M,F ) .
Introduction
51
16-Feb-17
Sisters Question
Question:
? - s i s t e r ( a l i c e , e d w a r d ) .
The question (goal) matches the head of the rule, if one
replaces X with a l i c e and Y with edward.
The instance of the body becomes new
goal: female(a l i c e ) ,
pa ren ts ( a l i c e , M , F ) ,
pa ren ts ( edward, M, F ) .
Introduction
52
16-Feb-17
7 s i s t e r ( a l i c e , e d wa r d )
X0=a l i c e , Y0=edward
3 f e m a l e ( a l i c e ) ,
p a r e n t s ( a l i c e ,M 0 , F0 ) ,
parents(edward,M0,F0).
6 p a r e n t s (a l i c e , M 0 , F0 ) ,
parents(edward,M0,F0).
M0 = v i c t o r i a , F0=alber t
5 p a r e n t s ( e d w a r d , v i c t o r i a , a l b e r t ) .
•
Is Alice Edward’s Sister?
( 1 ) m a l e ( a l b e r t ) .
(2) male(edward).
( 3 ) f e m a l e ( a l i c e ) .
( 4 ) f e m a l e ( v i c t o r i a ) .
(5) parents (edward ,
v i c t o r i a , a l b e r t ) .
( 6 ) p a r e n t s ( a l i c e ,
v i c t o r i a , a l b e r t ) .
( 7 ) s i s t e r ( X , Y ) : -
f em a le (X) ,
parents (X,M,F ) ,
pa ren t s (Y ,M,F ) .
Introduction
53
16-Feb-17
Is Alice Edward’s Sister?
7 s i s t e r ( a l i c e , e d wa r d )
X0=a l i c e , Y0=edward
3 f e m a l e ( a l i c e ) ,
p a r e n t s ( a l i c e ,M 0 , F0 ) ,
parents(edward,M0,F0).
6 p a r e n t s (a l i c e , M 0 , F0 ) ,
parents(edward,M0,F0).
M0 = v i c t o r i a , F0=alber t
5 p a r e n t s ( e d w a r d , v i c t o r i a , a l b e r t ) .
•
Introduction
54
16-Feb-17
( 1 ) m a l e ( a l b e r t ) .
(2) male(edward).
( 3 ) f e m a l e ( a l i c e ) .
( 4 ) f e m a l e ( v i c t o r i a ) .
(5) parents (edward ,
v i c t o r i a , a l b e r t ) .
( 6 ) p a r e n t s ( a l i c e ,
v i c t o r i a , a l b e r t ) .
( 7 ) s i s t e r ( X , Y ) : -
f em a le (X) ,
parents (X,M,F ) ,
pa ren t s (Y ,M,F ) .
Who’s Sister Is Alice?
7 s i s t e r ( a l i c e , X )
X0=a l i ce , Y0=X
3 f e m a l e ( a l i c e ) ,
p a r e n t s ( a l i c e ,M 0 , F0 ) ,
parents (X,M0,F0) .
6 p a r e n t s (a l i c e , M 0 , F0 ) ,
parents (X,M0,F0) .
M0=vic tor ia F0=alber t
5 p a r e n t s ( X , v i c t o r i a , a l b e r t ) .
X=edward
•
Answer: X = edward. Introduction
55
16-Feb-17
( 1 ) m a l e ( a l b e r t ) .
(2) male(edward).
( 3 ) f e m a l e ( a l i c e ) .
( 4 ) f e m a l e ( v i c t o r i a ) .
(5) parents (edward ,
v i c t o r i a , a l b e r t ) .
( 6 ) p a r e n t s ( a l i c e ,
v i c t o r i a , a l b e r t ) .
( 7 ) s i s t e r ( X , Y ) : -
f em a le (X) ,
parents (X,M,F ) ,
pa ren t s (Y ,M,F ) .
Useful Links
• SWI-Prolog:
• http://www.swi-prolog.org/
• SWI-Prolog Editor (Windows, SWI-x86 only)
http://lakk.bildung.hessen.de/netzwerk/faecher/informatik/swiprolog/ind
exe.html
• Prolog mode for (X)Emacs:
http://turing.ubishops.ca/home/bruda/emacs-prolog/
• Prolog newsgroup:
http://groups.google.com/groups?group=comp.lang.prolog
• Links from SWI-Prolog Web page:
http://www.swi-prolog.org/Links.html
Introduction
56
16-Feb-17