Optimization I (MAT 388)

Geometry of Linear Programming

Dipti Dubey

Department of Mathematics
Shiv Nadar University
CONVEX SET: A set S ⊆ Rn is said to be convex if for any two
points x, y in S and λ ∈ [0, 1],

λx + (1 − λ)y ∈ S.
CONVEX SET: A set S ⊆ Rn is said to be convex if for any two
points x, y in S and λ ∈ [0, 1],

λx + (1 − λ)y ∈ S.

• Note that λx + (1 − λ)y for λ ∈ [0, 1] represents a point on the

line segment joining x and y .
EXAMPLES:

1. Empty set and singleton sets are always considered to be convex.

2. Rn is convex.

3. S = {(x1 , x2 ) ∈ R2 | x12 + x22 ≤ 1} is a convex set.

4. S = {(x1 , x2 ) ∈ R2 | x12 + x22 = 1} is NOT a convex set

• Intersection of two convex sets is a convex set.

EXAMPLES:

1. Empty set and singleton sets are always considered to be convex.

2. Rn is convex.

3. S = {(x1 , x2 ) ∈ R2 | x12 + x22 ≤ 1} is a convex set.

4. S = {(x1 , x2 ) ∈ R2 | x12 + x22 = 1} is NOT a convex set

• Intersection of two convex sets is a convex set.

CONVEX COMBINATION: Any point of the form λx + (1 − λ)y

where λ ∈ [0, 1] is called a convex combination of points x and y .
EXAMPLES:

1. Empty set and singleton sets are always considered to be convex.

2. Rn is convex.

3. S = {(x1 , x2 ) ∈ R2 | x12 + x22 ≤ 1} is a convex set.

4. S = {(x1 , x2 ) ∈ R2 | x12 + x22 = 1} is NOT a convex set

• Intersection of two convex sets is a convex set.

CONVEX COMBINATION: Any point of the form λx + (1 − λ)y

where λ ∈ [0, 1] is called a convex combination of points x and y .

• If λ ∈ (0, 1), then the convex combination is called strict.

• We call a point of the form

λ1 x (1) + λ2 x (2) + . . . + λk x (k)

where x (i) ∈ Rn , λi ≥ 0 and ki=1 λi = 1, a convex combination
P
of x (1) , x (2) , . . . , x (k) .
• We call a point of the form

λ1 x (1) + λ2 x (2) + . . . + λk x (k)

where x (i) ∈ Rn , λi ≥ 0 and ki=1 λi = 1, a convex combination
P
of x (1) , x (2) , . . . , x (k) .

CONVEX HULL: The convex hull of a set S ⊆ Rn is the set of all

convex combinations of points in S, i.e.
k
X
Conv(S) = {λ1 x (1) +λ2 x (2) +. . .+λk x (k) | x (i) ∈ S, λi ≥ 0, λi = 1}
i=1
EXTREME POINT: A point x in a convex set S ⊆ Rn is said to
be an extreme point of S if x can not be represented as a strict
convex combination of two distinct points in S.
In other words, if

x = λx (1) + (1 − λ)x (2)

with λ ∈ (0, 1) and x (1) , x (2) ∈ S then x = x (1) = x (2) .
• z is an extreme point.
• Every point on the boundary is an extreme point.
Hyperplane: Let p ∈ Rn and k ∈ R.Then the set H defined as
H = {x ∈ Rn : p T x = k} is called a hyperplane.
Hyperplane: Let p ∈ Rn and k ∈ R.Then the set H defined as
H = {x ∈ Rn : p T x = k} is called a hyperplane.

Equivalently, hyperplane H consists of all the points x

(x T = (x1 , x2 , . . . , xn )) in Rn satisfying

p1 x1 + p2 x2 + . . . + pn xn = k.
Hyperplane: Let p ∈ Rn and k ∈ R.Then the set H defined as
H = {x ∈ Rn : p T x = k} is called a hyperplane.

Equivalently, hyperplane H consists of all the points x

(x T = (x1 , x2 , . . . , xn )) in Rn satisfying

p1 x1 + p2 x2 + . . . + pn xn = k.

• A hyperplane H in Rn is the natural extension of line in R2 or a

plane in R3 .
• Every hyperplane is a convex set.
A hyperplane divides the Rn into two regions, called half spaces.
A hyperplane divides the Rn into two regions, called half spaces.

Half Space: A half space is a collection of points of the form

{x ∈ Rn : p T x ≤ k}
or
{x ∈ Rn : p T x ≥ k}
where p is a nonzero vector in Rn and k ∈ R.
Theorem: Every half space is a convex set.

Proof. Consider a half space

S = {x ∈ Rn : p T x ≤ k}

where where p is a nonzero vector in Rn and k ∈ R.

Let x, y ∈ S and λ ∈ [0, 1] then
p T (λx + (1 − λ)y ) = λp T x + (1 − λ)p T y
≤ kλ + k(1 − λ)
= k.
This imples λx + (1 − λ)y ∈ S.
• Feasible set of a linear programming problem is a convex
set.
 • Feasible set of a linear programming problem is a convex
set.

Polyhedron:Let S ⊂ Rn . Then S is called a polyhedron if it is the

intersection of finite number of half spaces, i.e.

S = {x ∈ Rn : piT x ≤ ki , i = 1, . . . , r }.

Polytope: A bounded polyhedron is called a polytope.

RAYS AND DIRECTION:

RAY: A ray is collection of points of the form

{x0 + λd : λ ≥ 0},
where d is a non zero vector. Here x0 is called base (or vertex) of
the ray, and d is called direction of the ray.
RAYS AND DIRECTION:

RAY: A ray is collection of points of the form

{x0 + λd : λ ≥ 0},
where d is a non zero vector. Here x0 is called base (or vertex) of
the ray, and d is called direction of the ray.

DIRECTION: Given a convex set S, a non zero vector d is said

to be a direction of the set S, if for each x̂ ∈ S, the ray

{x̂ + λd : λ ≥ 0, }
also belongs to the set S.