Introduction to Modeling and Optimization

Graphical Solution of 2-variable LP Problems

Bilkent University IE202 Oya Karasan 1 / 65

Graphical Solution of 2-variable LP Problems

Ex 1.a)

max x1 + 3x2
s.t. x1 + x2 → 6 (1)
↑ x1 + 2x2 → 8 (2)
x1 , x2 ↓ 0 (3,4)

Bilkent University IE202 Oya Karasan 2 / 65

Graphical Solution of 2-variable LP Problems

Generalized form:
max x1 + 3x2 max cx
s.t. x1 + x2 → 6 (1) s.t. Ax → b
↑ x1 + 2x2 → 8 (2) x↓0
where,
x1 , x2 ↓ 0 (3,4) # $
! " x1
c= 1 3 x=
x2
# $ # $
1 1 6
A= b=
↑1 2 8

Bilkent University IE202 Oya Karasan 3 / 65

Bilkent University IE202 Oya Karasan 4 / 65
Bilkent University IE202 Oya Karasan 5 / 65
Bilkent University IE202 Oya Karasan 6 / 65
Bilkent University IE202 Oya Karasan 7 / 65
Graphical Solution of 2-variable LP Problems

A line on which all points have the same z-value (z = x1 + 3x2 ) is

called :

Isoprofit line for maximization problems,

Isocost line for minimization problems.

Bilkent University IE202 Oya Karasan 8 / 65

Graphical Solution of 2-variable LP Problems

Consider the coefficients of x1 and x2 in the

objective function and let c = [1 3].

Bilkent University IE202 Oya Karasan 9 / 65

Bilkent University IE202 Oya Karasan 10 / 65
Graphical Solution of 2-variable LP Problems

Note that as we move the isoprofit lines in the direction of c, the total
profit increases!

Bilkent University IE202 Oya Karasan 11 / 65

Graphical Solution of 2-variable LP Problems

For a maximization problem, the vector c, given by the coefficients of

x1 and x2 in the objective function, is called the improving direction.
On the other hand, -c is the improving direction for minimization
problems!

Bilkent University IE202 Oya Karasan 12 / 65

Graphical Solution of 2-variable LP Problems

To solve an LP problem in two variables, we should try to push isoprofit

(isocost) lines in the improving direction as much as possible while still
staying in the feasible region.

Bilkent University IE202 Oya Karasan 13 / 65

Bilkent University IE202 Oya Karasan 14 / 65
Bilkent University IE202 Oya Karasan 15 / 65
Bilkent University IE202 Oya Karasan 16 / 65
Bilkent University IE202 Oya Karasan 17 / 65
Bilkent University IE202 Oya Karasan 18 / 65
Bilkent University IE202 Oya Karasan 19 / 65
Bilkent University IE202 Oya Karasan 20 / 65
Bilkent University IE202 Oya Karasan 21 / 65
Graphical Solution of 2-variable LP Problems

Point (4/3, 14/3) is the last point in the feasible region when we move
in the improving direction. So, if we move in the same direction any
further, there is no feasible point.
Therefore, point z → = (4/3, 14/3) is the maximizing point. The optimal
value of the LP problem is 46/3.

Bilkent University IE202 Oya Karasan 22 / 65

Graphical Solution of 2-variable LP Problems

Let us modify the objective function while keeping the constraints

unchanged:

Ex 1.a) =↔ Ex 1.b)
max x1 + 3x2 max x1 ↑ 2x2
s.t. x1 + x2 → 6 (1) s.t. x1 + x2 → 6 (1)
↑ x1 + 2x2 → 8 (2) ↑ x1 + 2x2 → 8 (2)
x1 , x2 ↓ 0 (3,4) x1 , x2 ↓ 0 (3,4)

Bilkent University IE202 Oya Karasan 23 / 65

Graphical Solution of 2-variable LP Problems

Does feasible region change?

Bilkent University IE202 Oya Karasan 24 / 65

Bilkent University IE202 Oya Karasan 25 / 65
Graphical Solution of 2-variable LP Problems

What about the improving direction?

Bilkent University IE202 Oya Karasan 26 / 65

Bilkent University IE202 Oya Karasan 27 / 65
Bilkent University IE202 Oya Karasan 28 / 65
Graphical Solution of 2-variable LP Problems

Now let us consider:

Ex 1.c)
max ↑ x1 + x2
s.t. x1 + x2 → 6 (1)
↑ x1 + 2x2 → 8 (2)
x1 , x2 ↓ 0 (3,4)

Bilkent University IE202 Oya Karasan 29 / 65

Bilkent University IE202 Oya Karasan 30 / 65
Graphical Solution of 2-variable LP Problems

Now let us consider:

Question:
Ex 1.c) Is it possible for (5,1) be the
optimal?
max ↑ x1 + x2
s.t. x1 + x2 → 6 (1)
↑ x1 + 2x2 → 8 (2)
x1 , x2 ↓ 0 (3,4)

Bilkent University IE202 Oya Karasan 31 / 65

Bilkent University IE202 Oya Karasan 32 / 65
Bilkent University IE202 Oya Karasan 33 / 65
Bilkent University IE202 Oya Karasan 34 / 65
Graphical Solution of 2-variable LP Problems

Now let us consider:

Question:
Ex 1.d) Is it possible for (0,4) be the
optimal?
max ↑ x1 ↑ x2
s.t. x1 + x2 → 6 (1)
↑ x1 + 2x2 → 8 (2)
x1 , x2 ↓ 0 (3,4)

Bilkent University IE202 Oya Karasan 35 / 65

Bilkent University IE202 Oya Karasan 36 / 65
Graphical Solution of 2-variable LP Problems

Ex 2)
A company wishes to increase the demand for its product through
advertising. Each minute of radio ad costs $1 and each minute of TV
ad costs $2.
Each minute of radio ad increases the daily demand by 2 units and
each minute of TV ad by 7 units.
The company would wish to place at least 9 minutes of daily ad in
total. It wishes to increase daily demand by at least 28 units.
How can the company meet its advertising requirements at
minimum total cost?

Bilkent University IE202 Oya Karasan 37 / 65

Graphical Solution of 2-variable LP Problems

x1 : of minutes of radio ads purchased

x2 : of minutes of TV ads purchased

min x1 + 2x2
s.t. 2x1 + 7x2 ↓ 28 (1)
x1 + x2 ↓ 9 (2)
x1 , x2 ↓ 0 (3,4)

Bilkent University IE202 Oya Karasan 38 / 65

Bilkent University IE202 Oya Karasan 39 / 65
Bilkent University IE202 Oya Karasan 40 / 65
Bilkent University IE202 Oya Karasan 41 / 65
Bilkent University IE202 Oya Karasan 42 / 65
Graphical Solution of 2-variable LP Problems

Ex 3)
Suppose that in the previous example, it costed $4 to place a minute of
radio ad and $14 to place a minute of TV ad.

min 4x1 + 14x2

s.t. 2x1 + 7x2 ↓ 28 (1)
x1 + x2 ↓ 9 (2)
x1 , x2 ↓ 0 (3,4)

Bilkent University IE202 Oya Karasan 43 / 65

Bilkent University IE202 Oya Karasan 44 / 65
Bilkent University IE202 Oya Karasan 45 / 65
Graphical Solution of 2-variable LP Problems

# $ # $
7 ! " 7
At point A = , the objective function value is: 4 14 = 56.
2 2
# $ # $
14 ! " 14
At point B = , the objective function value is: 4 14 = 56.
0 0
This is also true for any feasible point on the line segment [A,B]. We
say that LP has multiple or alternate optimal solutions.
# $ # $
7 14
Line Segment [A, B] = {ω + (1 ↑ ω) : ω ↗ [0, 1]}
2 0

Bilkent University IE202 Oya Karasan 46 / 65

Graphical Solution of 2-variable LP Problems

Ex 4)
Consider the following LP problem:

max x1 + 3x2
s.t. ↑ x1 + 2x2 → 8 (1)
x1 , x2 ↓ 0 (2,3)

Bilkent University IE202 Oya Karasan 47 / 65

Bilkent University IE202 Oya Karasan 48 / 65
Unbounded LP Problems

The objective function for this maximization problem can be

increased by moving in the improving direction c as much as we want
while still staying in the feasible region. In this case, we say that the LP
is unbounded.
For unbounded LP problems, there is no optimal solution and the
optimal value is defined to be +↘.

Bilkent University IE202 Oya Karasan 49 / 65

Unbounded LP Problems

For the minimization problem, we say that the LP is unbounded if we

can DECREASE the objective function value as much as we want
while still staying in the feasible region.
In this case, the optimal value is defined to be -↘.

Bilkent University IE202 Oya Karasan 50 / 65

Graphical Solution of 2-variable LP Problems

Ex 5)

max x1 + 3x2
s.t. x1 + x2 → 6 (1)
↑ x1 + 2x2 → 8 (2)
x2 ↓ 6 (3)
x1 , x2 ↓ 0 (4,5)

Bilkent University IE202 Oya Karasan 51 / 65

Bilkent University IE202 Oya Karasan 52 / 65
Every LP Problem falls into one of the 4 cases:

Case 1: The LP problem has a unique optimal solution (see Ex.1 and
Ex.2)
Case 2: The LP problem has alternative or multiple optimal solution
(see Ex.3). In this case, there are infinitely many optimal solutions.
Case 3: The LP problem is unbounded (see Ex.4). In this case, there
is no optimal solution.
Case 4: The LP problem is infeasible (see Ex.5). In this case, there is
no feasible solution.

Bilkent University IE202 Oya Karasan 53 / 65

Graphical Solution of LP Problems:

We may graphically solve an LP with two decision variables as follows:

Step 1: Graph the feasible region.
Step 2: Draw an isoprofit line (for max problem) or an isocost line (for
min problem).
Step 3: Move parallel to the isoprofit/isocost line in the improving
direction. The last point in the feasible region that contacts an
isoprofit/isocost line is an optimal solution to the LP.

Bilkent University IE202 Oya Karasan 54 / 65

Graphical Solution of 2-variable LP Problems

Have you noticed anything about the optimal points?

Bilkent University IE202 Oya Karasan 55 / 65

Convex Set:

The convex combination of two points x and y is given by the point

ωx + (1 ↑ ω)y for any ω ↗ [0, 1]
A set of points S is a convex set if for any two points x ↗ S and y ↗ S,
their convex combination ωx + (1 ↑ ω)y is also in S for all ω ↗ [0, 1].
In other words, a set is a convex set if the line segment joining any pair
of points in S is wholly contained in S.

Bilkent University IE202 Oya Karasan 56 / 65

Are those Convex Sets?

Bilkent University IE202 Oya Karasan 57 / 65

Are those Convex Sets?

Feasible set of an LP x = {x ↗ R n : Ax → b, x ↓ 0} is a convex set.

Proof: Exercise

Bilkent University IE202 Oya Karasan 58 / 65

Extreme Points (Corner Points):

A point P of a convex set S is an extreme point (corner point) if it

cannot be represented as a strict convex combination of two distinct
points of S.
strict convex combination: x is a strict convex combination of x1 and
x2 if x = ωx1 + (1 ↑ ω)x2 for some ω ↗ (0, 1).

Bilkent University IE202 Oya Karasan 59 / 65

Extreme Points (Corner Points):

In other words, for any convex set S, a point P in S is an extreme point

(corner point) if each line segment that lies completely in S and
contains the point P, has P as an endpoint of the line segment.

Bilkent University IE202 Oya Karasan 60 / 65

Fundamental Theorem of LP:

1) If the feasible set {x ↗ R n : Ax → b, x ↓ 0} is not empty, that is if

there is a feasible solution, then there is an extreme point.
2) If an LP has an optimal solution, then there is an extreme (or corner)
point of the feasible region which is an optimal solution to the LP
P.S: Not every optimal solution needs to be an extreme point!
(Remember the alternate optimal solution case)

Bilkent University IE202 Oya Karasan 61 / 65

Remarks

1) There exists an objective function such that every corner point x is

the (unique) optimal solution. This happens when the improving
direction is (strictly) in the cone of exit at x

Bilkent University IE202 Oya Karasan 62 / 65

Bilkent University IE202 Oya Karasan 63 / 65
Remarks

1) There exists an objective function such that every corner point x is

the (unique) optimal solution. This happens when the improving
direction is (strictly) in the cone of exit at x
2) In order to have the LP unbounded, FR must be unbounded,
however, the reverse is not necessarily true

Bilkent University IE202 Oya Karasan 64 / 65

Bilkent University IE202 Oya Karasan 65 / 65