Forward and Backword Chaining problems:

Example 1:
Given Facts (Knowledge Base):
1. John owns a dog (John → owns(dog)).
2. Buddy is a dog (Buddy → dog).
3. Rule: If a person owns a dog, they love animals (If owns(person, dog),
then loves(person, animals)).
Goal:
Prove that John loves animals.

Forward Chaining (Data-driven approach):

1. Start with known facts:
o John owns a dog (John → owns(dog)).

o The rule tells us that if someone owns a dog, they love animals.

2. Apply the rule:

o Since John owns a dog, we can apply the rule: "If a person owns a
dog, they love animals."
3. Conclusion:
o Therefore, John loves animals.

Backward Chaining (Goal-driven approach):

1. Start with the goal:
o The goal is to prove that John loves animals.

2. Check the rule:

o Look at the rule: "If a person owns a dog, they love animals." To
prove that John loves animals, we need to verify if John owns a dog.
3. Check the fact:
o The fact tells us that John owns a dog.

4. Conclusion:
o Since John owns a dog, the rule applies, and we can conclude that
John loves animals.
Example 2:
Given Facts (Knowledge Base):
1. Alice owns a cat.
2. Fluffy is a cat.
3. Rule: If a person owns a cat, they care for cats.
4. Goal: Prove Alice cares for cats.
Forward Chaining (Data-driven approach):
1. Start with known facts:
o Alice owns a cat (Alice → owns(cat)).

o The rule tells us that if someone owns a cat, they care for cats.

2. Apply the rule:

o Since Alice owns a cat, apply the rule: "If a person owns a cat, they
care for cats."
3. Conclusion:
o Therefore, Alice cares for cats.

Backward Chaining (Goal-driven approach):

1. Start with the goal:
o The goal is to prove that Alice cares for cats.

2. Check the rule:

o Look at the rule: "If a person owns a cat, they care for cats." To
prove that Alice cares for cats, we need to verify if Alice owns a cat.
3. Check the fact:
o The fact tells us that Alice owns a cat.

4. Conclusion:
o Since Alice owns a cat, the rule applies, and we conclude that Alice
cares for cats.
Example 3:
Given Facts (Knowledge Base):
1. Tom owns a hamster.
2. Hammy is a hamster.
3. Rule: If a person owns a hamster, they clean the cage.
4. Goal: Prove Tom cleans the cage.
Forward Chaining (Data-driven approach):
 1. Start with known facts:
o Tom owns a hamster (Tom → owns(hamster)).

o The rule tells us that if someone owns a hamster, they clean the
cage.
2. Apply the rule:
o Since Tom owns a hamster, apply the rule: "If a person owns a
hamster, they clean the cage."
3. Conclusion:
o Therefore, Tom cleans the cage.

Backward Chaining (Goal-driven approach):

1. Start with the goal:
o The goal is to prove that Tom cleans the cage.

2. Check the rule:

o Look at the rule: "If a person owns a hamster, they clean the cage."
To prove that Tom cleans the cage, we need to verify if Tom owns a
hamster.
3. Check the fact:
o The fact tells us that Tom owns a hamster.

4. Conclusion:
o Since Tom owns a hamster, the rule applies, and we conclude that
Tom cleans the cage.

Resolution Problems :
Example 1:
Given Facts (Knowledge Base):
1. All birds can fly.
2. Penguins are birds.
3. Some birds are penguins.
4. Tweety is a bird.
Goal: Prove that some birds cannot fly.
Step-by-step Process using Resolution:
1. Convert the facts into CNF:

∀x(Bird(x)→CanFly(x))\forall x (Bird(x) \rightarrow

o Fact 1 (All birds can fly):

CanFly(x))∀x(Bird(x)→CanFly(x))
 In CNF:
¬Bird(x)∨CanFly(x)\neg Bird(x) \lor CanFly(x)¬Bird(x)∨CanFly(x)

∀x(Penguin(x)→Bird(x))\forall x (Penguin(x) \rightarrow

o Fact 2 (Penguins are birds):

Bird(x))∀x(Penguin(x)→Bird(x))
In CNF:
¬Penguin(x)∨Bird(x)\neg Penguin(x) \lor Bird(x)¬Penguin(x)∨Bird(x)

∃x(Penguin(x))\exists x (Penguin(x))∃x(Penguin(x))
o Fact 3 (Some birds are penguins):

In CNF:
Penguin(a)Penguin(a)Penguin(a) (where a is a constant denoting a
penguin)
o Fact 4 (Tweety is a bird):
Bird(Tweety)Bird(Tweety)Bird(Tweety)
2. Negate the goal:
The goal is to prove that some birds cannot fly.

∀x(Bird(x)→CanFly(x))\forall x (Bird(x) \rightarrow

We negate it:

CanFly(x))∀x(Bird(x)→CanFly(x))
This is equivalent to:
¬Bird(x)∨CanFly(x)\neg Bird(x) \lor CanFly(x)¬Bird(x)∨CanFly(x) (which is
already in CNF form).
Now, we need to prove this negation leads to a contradiction.
3. Resolution:
Now we combine the clauses:

o From Fact 2: ¬Penguin(x)∨Bird(x)\neg Penguin(x) \lor

Bird(x)¬Penguin(x)∨Bird(x)

o From Fact 1: ¬Bird(x)∨CanFly(x)\neg Bird(x) \lor

CanFly(x)¬Bird(x)∨CanFly(x)
o From Fact 3: Penguin(a)Penguin(a)Penguin(a)

o From Fact 4: Bird(Tweety)Bird(Tweety)Bird(Tweety)

We resolve the clauses step by step:

o From Fact 3 and Fact 2, we resolve on
Penguin(a)Penguin(a)Penguin(a) to get:
Bird(a)Bird(a)Bird(a)
o Using Bird(a) in Fact 1, we resolve to get:
CanFly(a)CanFly(a)CanFly(a)
But this is a contradiction to our goal. Since we have derived CanFly(a), and a is
a penguin, this shows that some birds (like penguins) cannot fly, which proves
the goal.
Example 2:
Given Facts (Knowledge Base):
1. All dogs bark.
2. All animals that bark are mammals.
3. Fido is a dog.
4. Some animals are not mammals.
Goal: Prove that Fido is not a mammal.
Step-by-step Process using Resolution:
1. Convert the facts into CNF:

∀x(Dog(x)→Barks(x))\forall x (Dog(x) \rightarrow

o Fact 1 (All dogs bark):

Barks(x))∀x(Dog(x)→Barks(x))
In CNF:
¬Dog(x)∨Barks(x)\neg Dog(x) \lor Barks(x)¬Dog(x)∨Barks(x)

∀x(Barks(x)→Mammal(x))\forall x (Barks(x) \rightarrow

o Fact 2 (All animals that bark are mammals):

Mammal(x))∀x(Barks(x)→Mammal(x))
In CNF:
¬Barks(x)∨Mammal(x)\neg Barks(x) \lor
Mammal(x)¬Barks(x)∨Mammal(x)
o Fact 3 (Fido is a dog):
Dog(Fido)Dog(Fido)Dog(Fido)

∃x(¬Mammal(x))\exists x (\neg Mammal(x))∃x(¬Mammal(x))

o Fact 4 (Some animals are not mammals):

In CNF:
¬Mammal(a)\neg Mammal(a)¬Mammal(a) (where a is a constant
denoting a non-mammal)
2. Negate the goal:
The goal is to prove that Fido is not a mammal.
We negate it:
Mammal(Fido)Mammal(Fido)Mammal(Fido)
3. Resolution:
Now, we combine the clauses:
o From Fact 3: Dog(Fido)Dog(Fido)Dog(Fido)
 o From Fact 1: ¬Dog(x)∨Barks(x)\neg Dog(x) \lor
Barks(x)¬Dog(x)∨Barks(x), which gives
Barks(Fido)Barks(Fido)Barks(Fido)

o From Fact 2: ¬Barks(x)∨Mammal(x)\neg Barks(x) \lor

Mammal(x)¬Barks(x)∨Mammal(x), which gives
Mammal(Fido)Mammal(Fido)Mammal(Fido)
But the negation of the goal says ¬Mammal(Fido)\neg
Mammal(Fido)¬Mammal(Fido), which leads to a contradiction with
Mammal(Fido)Mammal(Fido)Mammal(Fido).
Therefore, Fido is not a mammal, as we set out to prove.

Example 3:
Given Facts (Knowledge Base):
1. No humans are aliens.
2. All humans are mortal.
3. John is a human.
4. Some humans are philosophers.
Goal: Prove that no philosophers are aliens.
Step-by-step Process using Resolution:
1. Convert the facts into CNF:

∀x(Human(x)→¬Alien(x))\forall x (Human(x) \rightarrow \neg

o Fact 1 (No humans are aliens):

Alien(x))∀x(Human(x)→¬Alien(x))
In CNF:
¬Human(x)∨¬Alien(x)\neg Human(x) \lor \neg
Alien(x)¬Human(x)∨¬Alien(x)

∀x(Human(x)→Mortal(x))\forall x (Human(x) \rightarrow

o Fact 2 (All humans are mortal):

Mortal(x))∀x(Human(x)→Mortal(x))
In CNF:
¬Human(x)∨Mortal(x)\neg Human(x) \lor
Mortal(x)¬Human(x)∨Mortal(x)
o Fact 3 (John is a human):
Human(John)Human(John)Human(John)

∃x(Philosopher(x))\exists x (Philosopher(x))∃x(Philosopher(x))
o Fact 4 (Some humans are philosophers):

In CNF:
 Philosopher(a)Philosopher(a)Philosopher(a) (where a is a constant
denoting a philosopher)
2. Negate the goal:
The goal is to prove that no philosophers are aliens.

∃x(Philosopher(x)∧Alien(x))\exists x (Philosopher(x) \land

We negate it:

Alien(x))∃x(Philosopher(x)∧Alien(x))
In CNF:
Philosopher(a)∧Alien(a)Philosopher(a) \land
Alien(a)Philosopher(a)∧Alien(a)
3. Resolution:
Now, we combine the clauses:
o From Fact 4: Philosopher(a)Philosopher(a)Philosopher(a)

o From Fact 1: ¬Human(x)∨¬Alien(x)\neg Human(x) \lor \neg

Alien(x)¬Human(x)∨¬Alien(x), which gives ¬Alien(a)\neg
Alien(a)¬Alien(a)
Resolving these clauses, we find a contradiction. Since Alien(a) cannot be true,
we conclude that no philosophers are aliens, as required.

Finding Most General Unifier (MGU)

Example 1:
Find the MGU of the following terms:
 g(f(X),Y)g(f(X), Y)g(f(X),Y) and g(f(a),b)g(f(a), b)g(f(a),b)
Step-by-step solution:
1. Compare the two terms:
Both are functions g(…)g(\ldots)g(…). Let's compare the arguments:
o First argument: f(X)f(X)f(X) and f(a)f(a)f(a)

o Second argument: YYY and bbb

2. Unify the arguments one by one:

o First argument:
f(X)f(X)f(X) and f(a)f(a)f(a)
 Both are functions f(…)f(\ldots)f(…), so we compare their
inner arguments:
 We need to unify XXX and aaa, so we substitute XXX with
aaa.

 So, X↦aX \mapsto aX↦a.

 o Second argument:
YYY and bbb
 YYY is a variable, and bbb is a constant. To make them equal,
we substitute YYY with bbb.

 So, Y↦bY \mapsto bY↦b.

Final Substitutions:

 X↦aX \mapsto aX↦a

 Y↦bY \mapsto bY↦b

MGU:
The Most General Unifier (MGU) is:

{X↦a,Y↦b}\{ X \mapsto a, Y \mapsto b \}{X↦a,Y↦b}

Example 2:
To find the Most General Unifier (MGU) for the two terms:
 J(b,X,f(g(Z)))J(b, X, f(g(Z)))J(b,X,f(g(Z)))
 J(Z,f(P),f(P))J(Z, f(P), f(P))J(Z,f(P),f(P))
Step-by-step procedure:
1. Compare the two terms:
Both expressions are predicates of the form J(…)J(\ldots)J(…), meaning the
functor JJJ is the same for both. We will compare the arguments of the predicates
position by position:
o Argument 1: bbb and ZZZ

o Argument 2: XXX and f(P)f(P)f(P)

o Argument 3: f(g(Z))f(g(Z))f(g(Z)) and f(P)f(P)f(P)

2. Start unifying the arguments one by one:

o First argument:
bbb and ZZZ
 Here, bbb is a constant, and ZZZ is a variable. To make these
two equal, we substitute ZZZ with bbb.

 So, Z↦bZ \mapsto bZ↦b.

o Second argument:
XXX and f(P)f(P)f(P)
 Here, XXX is a variable and f(P)f(P)f(P) is a function. To make
them equal, we must substitute XXX with f(P)f(P)f(P).

 So, X↦f(P)X \mapsto f(P)X↦f(P).

 o Third argument:
f(g(Z))f(g(Z))f(g(Z)) and f(P)f(P)f(P)
 f(g(Z))f(g(Z))f(g(Z)) is a function, and f(P)f(P)f(P) is another
function. To unify them, the arguments inside the functions
must also be unified.
 So, we need to unify g(Z)g(Z)g(Z) and PPP.

 From the earlier unification step, we know that Z↦bZ \mapsto

bZ↦b. Therefore, we need to unify g(b)g(b)g(b) and PPP.
 To unify these, we substitute PPP with g(b)g(b)g(b).

 So, P↦g(b)P \mapsto g(b)P↦g(b).

Final Substitutions:

1. Z↦bZ \mapsto bZ↦b

2. X↦f(P)X \mapsto f(P)X↦f(P)

3. P↦g(b)P \mapsto g(b)P↦g(b)

MGU:
The Most General Unifier (MGU) is:

{Z↦b,X↦f(g(b)),P↦g(b)}\{ Z \mapsto b, X \mapsto f(g(b)), P \mapsto g(b) \}

{Z↦b,X↦f(g(b)),P↦g(b)}
This is the unification of the two terms J(b,X,f(g(Z)))J(b, X, f(g(Z)))J(b,X,f(g(Z)))
and J(Z,f(P),f(P))J(Z, f(P), f(P))J(Z,f(P),f(P)).