Solving Linear

Programming Problems
The Simplex Method (PART 2 )

Mohammed Brahimi
ENSIA/Intelligent Systems Enginnering

October 19, 2024

Outline
Key Takeaways from Last Lecture

Questions Regarding Simplex

Two-Phase method

Simplex: Special cases

Infeasibility

Unboundedness

Alternative Optima

Degeneracy and cycling

Sensitivity Analysis

2/39
Key Takeaways from Last Lecture
• Simplex is an efficient algorithm for finding optimal solutions to LP problems by
navigating through the corner points of the feasible region.

• It iteratively moves from one Basic Feasible Solution (BFS) to a better neighborhood
BFS until the optimal BFS is reached.

• By detecting the optimal BFS, the simplex method provides the optimal values of
the decision variables and the objective function.

3/39
Questions Regarding Simplex
• How can we choose an appropriate initial BFS if the origin is not a basic feasible
solution (BFS)?

4/39
Questions Regarding Simplex
• How can we choose an appropriate initial BFS if the origin is not a basic feasible
solution (BFS)?

• What are the special cases that may arise when using Simplex?

4/39
Questions Regarding Simplex
• How can we choose an appropriate initial BFS if the origin is not a basic feasible
solution (BFS)?

• What are the special cases that may arise when using Simplex?

• Does Simplex terminate in every LP?

4/39
Artificial starting solution
• LPs where all constraints are of the form ”≤” with nonnegative right-hand sides can
be conveniently started with an all-slack basic feasible solution.
• Select ”all-slack” variables as basic variables to create an initial basic feasible
solution (BFS).

∑
n

S1 = b1 + a1i xi
i=m+1

..
.
∑
n

Sm = bm + ami xi
i=m+1

5/39
Artificial starting solution
• If LPs involves constraints of the form ”≥” or ”=” do not have this convenient
starting solution.
• This ”ill-behaved” LPs, artificial variables should be used to find an initial BFS that
we can start with.

Maximize 3x + 9 y
Subject to x+y ≤3
5x − y ≥ 3
y ≥1

6/39
Two-Phase method
Maximize c1 x1 + c2 x2 + · · · + cn xn
Subject to a11 x1 + a12 x2 + · · · + a1n xn = b1
a21 x1 + a22 x2 + · · · + a2n xn = b2
..
.
am1 x1 + am2 x2 + · · · + amn xn = bm
x1 , x2 , . . . , xn ≥0

7/39
Two-Phase method
Minimize R1 + R2 + · · · + Rm
Subject to a11 x1 + a12 x2 + · · · + a1n xn + R1 = b1
a21 x1 + a22 x2 + · · · + a2n xn + R2 = b2
..
.
am1 x1 + am2 x2 + · · · + amn xn + Rm = bm
• Feasible if the objective value reaches 0.
• All the Ri are zeros.
• I have a BFS without the Ri .

7/39
Example: The Standard form (Phase 1)

Maximize z = 3 x + 9y
Subject to x + y ≤ 3
5 x − y ≥ 3
y ≥ 1

8/39
Example: The Standard form (Phase 1)

Maximize z = 3 x + 9y
Subject to x + y + S1 = 3
5 x − y − S2 = 3
y − S3 = 1

8/39
Example: Two-Phase method (Phase 1)

Minimize z = R1 + R2 + R3
Subject to x + y + S1 + R1 = 3
5x − y − S2 + R2 = 3
y − S3 + R3 = 1

9/39
Example: Two-Phase method (Phase 1)
Basic x y S1 S2 S3 R1 R2 R3 RHS Ratio
z 0 0 0 0 0 -1 -1 -1 0
R1 1 1 1 0 0 1 0 0 3
R2 5 -1 0 -1 0 0 1 0 3
R3 0 1 0 0 -1 0 0 1 1

• Basic variables = {R1 , R2 , R3 }

• We should remove basic variables from the objective function to start Simplex.
• Row(z) = Row(z) + Row(R1 ) + Row(R2 ) + Row(R3 )

9/39
Example: Two-Phase method (Phase 1)
Basic x y S1 S2 S3 R1 R2 R3 RHS Ratio
z 6 1 1 -1 -1 0 0 0 7
R1 1 1 1 0 0 1 0 0 3
R2 5 -1 0 -1 0 0 1 0 3
R3 0 1 0 0 -1 0 0 1 1

• Entering variable: S1 Leaving Variable: R1

• Basic variables = {S1 , R2 , R3 }
• We should remove basic variables from the objective function to start Simplex.
• Row(z) = Row(z) − Pivot row

9/39
Example: Two-Phase method (Phase 1)
Basic x y S1 S2 S3 R1 R2 R3 RHS Ratio
z 5 0 0 -1 -1 -1 0 0 4
S1 1 1 1 0 0 1 0 0 3 3
R2 5 -1 0 -1 0 0 1 0 3 3/5
R3 0 1 0 0 -1 0 0 1 1

• Entering variable: x, Leaving Variable: R2

• Basic variables = {S1 , x, R3 }
• Row(z) = Row(z) − 5*Pivot row
• Row(S1 ) = Row(S1 ) − Pivot row

9/39
Example: Two-Phase method (Phase 1)
Basic x y S1 S2 S3 R1 R2 R3 RHS Ratio
z 0 1 0 0 -1 -1 -1 0 1
S1 0 6/5 1 1/5 0 1 -1/5 0 12/5 2
x 1 -1/5 0 -1/5 0 0 1/5 0 3/5
R3 0 1 0 0 -1 0 0 1 1 1

• Entering variable: y, Leaving Variable: R3

• Basic variables = {S1 , x, y}
• Row(z) = Row(z) − Pivot row
• Row(S1 ) = Row(S1 ) − 65 Pivot row
• Row(x) = Row(x) + 15 Pivot row

9/39
Example: Two-Phase method (Phase 1)
Basic x y S1 S2 S3 R1 R2 R3 RHS Ratio
z 0 0 0 0 0 -1 -1 -1 0
S1 0 0 1 1/5 6/5 1 - 1/5 -6/5 6/5
x 1 0 0 -1/5 -1/5 0 1/5 1/5 4/5
y 0 1 0 0 -1 0 0 1 1

• Optimally detected because (Ci ≤ 0, ∀i).

• R1 = R2 = R3 = 0
• Basic variables = {S1 , x, y}
• Non-basic variables = {S2 , S3 }

9/39
Example: Two-Phase method (Phase 2)

Maximize z = 3 x + 9 y
6 1 6
Subject to S1 = + S2 + S3
5 5 5
4 1 1
x = − S2 − S3
5 5 5
y = 1 − S3

10/39
 3
x + y <= 3
5x - y >= 3
y >= 1

2
y

0
0 1 2 3
x

11/39
Example: Two-Phase method (Phase 2)
Basic x y S1 S2 S3 RHS Ratio
z -3 -9 0 0 0 0
S1 0 0 1 1/5 6/5 6/5
x 1 0 0 -1/5 -1/5 4/5
y 0 1 0 0 -1 1

• Basic variables = {S1 , x, y}

• We should remove basic variables from the objective function to start Simplex.
• Row(z) = Row(z) + 3 × Row(x) + 9 × Row(y)

12/39
Example: Two-Phase method (Phase 2)
Basic x y S1 S2 S3 RHS Ratio
z 0.00 0.00 0.00 -3/5 -48/5 57/5
S1 0.00 0.00 1.00 1/5 6/5 6/5
x 1 0 0 -1/5 -1/5 4/5
y 0 1 0 0 -1 1

• Entering variable: S3 , Leaving Variable: S1

• Pivot row = 5
6
× Pivot row
• Row(z) = Row(z) + 48
5
× Pivot row
• Row(x) = Row(x) + 1
5
× Pivot row
• Row(y) = Row(y) + Pivot row

12/39
Example: Two-Phase method (Phase 2)
Basic x y S1 S2 S3 RHS Ratio
z 0.00 0.00 8.00 1.00 0.00 21.00
S3 0 0 5/6 1/6 1 1
x 1 0 1/6 -1/6 0 1
y 0 1 5/6 1/6 0 2

• Optimally detected because (Ci ≥ 0, ∀i).

• Basic variables = {S3 , x, y}
• Non-basic variables = {S1 , S2 }
• Optimum (x = 1, y = 2) and z = 21.

12/39
 3
x + y <= 3
5x - y >= 3
y >= 1

2
y

0
0 1 2 3
x

13/39
Simplex: Special cases
• Infeasibility: occurs when there is no feasible solution that satisfies all of the
constraints.

• Unboundedness: occurs when the objective function can be increased indefinitely

without violating any of the constraints.

• Alternative Optima: occurs when several global optima with same objective value
exists.

• Degeneracy: occurs when one or more basic variables become zero during the
iteration process.

14/39
 Special cases:Infeasibility
• Empty feasible region.
Maximize z = 2x1 − 3x2
• Can be detected using
S.t. x1 + x2 ≤ 2
Two-Phase method.
∑ 2x1 − 2x2 ≥ 5
• The objective function ( Ri ) in x1 , x2 ≥ 0
Phase 1 cannot be 0.

15/39
Special cases:Unboundedness
• Unbounded solutions allow for arbitrary increases in variables without violating
any constraints.

• Unboundedness may indicate a poorly constructed model.

Maximize z = 2 x1 + x2
S.t x1 − x2 ≤ 10
2 x1 ≤ 40
x1 , x2 ≥ 0

16/39
 Special cases:Unboundedness

Maximize z = 2 x1 + x2
S.t x1 − x2 ≤ 10
2 x1 ≤ 40
x1 , x2 ≥ 0

17/39
How to detect Unboundedness
Simplex Method indicates unbounded solutions when all Ratios values
are either infinite or negative, resulting in no leaving variable.

Basic x1 x2 S1 S2 RHS Ratio

z -2 -1 0 0 0
S1 1 -1 1 0 10 Negative
S2 2 0 0 1 40 Infinite

18/39
Special cases: Alternative Optima
• An LP problem may have infinite alternative optima when the objective function is
parallel to a constraint.

• Any point on that constraint line is also optimal.

• Alternative optima provide different variable combinations with the same optimal
objective value.

Maximize z = 2 x1 + 4 x2
S.t x1 + 2 x2 ≤ 5
x1 + x2 ≤ 4
x1 , x2 ≥ 0

19/39
 Special cases: Alternative Optima

Max z =2 x1 + 4x2
S.t x1 + 2x2 ≤5
x1 + x2 ≤4
x1 , x2 ≥0

20/39
Special cases: Alternative Optima

Basic x1 x2 S1 S2 RHS
z -2 -4 0 0 0
S1 1 2 1 0 5
S2 1 1 0 1 4

• Entering variable: x2 , Leaving Variable: S1

• Pivot row = 1
2
× Pivot row
• Row(z) = Row(z) + 4 × Pivot row
• Row(S2 ) = Row(S2 ) − Pivot row

21/39
Special cases: Alternative Optima

• If a non-basic variable has a zero coefficient, it can be replaced with a basic

variable whose right-hand side value is strictly positive without changing the
objective function’s right-hand side value.
• If we swap x1 with x2 , where x2 has a right-hand side value of 2.5, then the
objective function’s right-hand side value remains at 10. After the swap, x1
becomes the new basic variable with a value of 5, and x2 becomes a non-basic
variable with a value of 0.

Basic x1 x2 S1 S2 RHS
z 0 0 2 0 10
x2 0.5 1 0.5 0 2.5
S2 0.5 0 -0.5 1 1.5

21/39
Special cases: Degeneracy
• Feasibility condition of simplex method can have ties for minimum ratio.

• Ties can be broken arbitrarily but will result in a degenerate solution in the next
iteration

• Degeneracy can cause the algorithm to cycle indefinitely and not terminate

Maximize z = 3 x1 + 9 x2
S.t x1 + 4 x2 ≤ 8
x1 + 2 x2 ≤ 4
x1 , x2 ≥ 0

22/39
 Special cases: Degeneracy

Max z =3 x1 + 9x2
S.t x1 + 4x2 ≤8
x1 + 2x2 ≤4
x1 , x2 ≥0

23/39
Special cases: Degeneracy

• Ties in minimum ratio.

• Some basic variable are equal to zero.

Basic x1 x2 S1 S2 RHS Ratio

z -3 -9 0 0 0
S1 1 4 1 0 8 2
S2 1 2 0 1 4 2

24/39
Special cases: Degeneracy

• Ties in minimum ratio.

• Some basic variable are equal to zero.

Basic x1 x2 S1 S2 RHS Ratio

z -0.75 0 2.25 0 18
x2 0.25 1 0.25 0 2 8
S2 0.5 0 -0.5 1 0 0

24/39
Special cases: Degeneracy

• Ties in minimum ratio.

• Some basic variable are equal to zero.

Basic x1 x2 S1 S2 RHS Ratio

z 0 0 1.5 1.5 18
x2 0 1 0.5 -0.5 2
x1 1 0 -1 2 0

24/39
Degeneracy interpretation

• The presence of degeneracy in an LP suggests the potential existence of a

superfluous constraint.

• Shuffling around the basic variables without departing from a corner.

• Dealing with degeneracy in an LP can create the impression that we are moving
from one corner to another, while keeping the objective value constant.

25/39
Degeneracy can cause cycling
• Cycling happens when the simplex algorithm loops between multiple solutions
without reaching the optimal solution due to degeneracy.

• This can cause the simplex algorithm to loop indefinitely.

• To prevent cycling, anti-cycling rules, such as Bland’s rule, can be applied to stop
revisiting the same solution and improve the efficiency of the simplex algorithm.

• If there are multiple ratios that are minimal, choose the variable xj with the
smallest index as the entering variable.

26/39
Example of cycling

Max z =2.3 x1 + 2.15x2 −13.55 x3 − 0.4x4

S.t 0.4 x1 + 0.2x2 −1.4 x3 − 0.2x4 ≤0
− 7.8 x1 − 1.4x2 +7.8 x3 + 0.4x4 ≤0
x1 , x2 x3 , x4 ≥0

27/39
Example of cycling

Basic x1 x2 x3 x4 x5 x6 RHS
z -2.3 -2.15 13.55 0.4 0 0 0
x5 0.4 0.2 -1.4 -0.2 1 0 0
x6 -7.8 -1.4 7.8 0.4 0 1 0

We are at the origin (0, 0, ...0)

28/39
Example of cycling

Basic x1 x2 x3 x4 x5 x6 RHS
z 0 -1 5.5 -0.75 5.75 0 0
x1 1 0.5 -3.5 -0.5 2.5 0 0
x6 0 2.5 -19.5 -3.5 19.5 1 0

We are at the origin (0, 0, ...0)

28/39
Example of cycling

Basic x1 x2 x3 x4 x5 x6 RHS
z 0 0 -2.3 -2.15 13.55 0.4 0
x1 1 0 0.4 0.2 -1.4 -0.2 0
x2 0 1 -7.8 -1.4 7.8 0.4 0

We are at the origin (0, 0, ...0).

28/39
Example of cycling

Basic x1 x2 x3 x4 x5 x6 RHS
z 5.75 0 0 -1 5.5 -0.75 0
x3 2.5 0 1 0.5 -3.5 -0.5 0
x6 19.5 1 0 2.5 -19.5 -3.5 0

We are at the origin (0, 0, ...0).

28/39
Example of cycling

Basic x1 x2 x3 x4 x5 x6 RHS
z 13.55 0.4 0 0 -2.3 -2.15 0
x3 -1.4 -0.2 1 0 0.4 0.2 0
x4 7.8 0.4 0 1 -7.8 -1.4 0

We are at the origin (0, 0, ...0).

28/39
Example of cycling

Basic x1 x2 x3 x4 x5 x6 RHS
z 5.5 -0.75 5.75 0 0 -1 0
x5 -3.5 -0.5 2.5 0 1 0.5 0
x4 -19.5 -3.5 19.5 1 0 2.5 0

We are at the origin (0, 0, ...0).

28/39
Example of cycling

Basic x1 x2 x3 x4 x5 x6 RHS
z -2.3 -2.15 13.55 0.4 0 0 0
x5 0.4 0.2 -1.4 -0.2 1 0 0
x4 -7.8 -1.4 7.8 0.4 0 1 0

The Simplex has returned to its original state.

28/39
Example of cycling
Basic x1 x2 x3 x4 x5 x6 RHS
z -2.3 -2.15 13.55 0.4 0 0 0
x5 0.4 0.2 -1.4 -0.2 1 0 0
x6 -7.8 -1.4 7.8 0.4 0 1 0

Basic x1 x2 x3 x4 x5 x6 RHS
z -2.3 -2.15 13.55 0.4 0 0 0
x5 0.4 0.2 -1.4 -0.2 1 0 0
x4 -7.8 -1.4 7.8 0.4 0 1 0

The Simplex will continuously cycle through these states.

29/39
Sensitivity Analysis
• Sensitivity analysis (or post-optimality analysis) determines how optimal solutions
are affected by changes within specified ranges.
– Changes in right-hand side (RHS) values.
– Changes in objective function coefficients.

30/39
Sensitivity Analysis
• Sensitivity analysis (or post-optimality analysis) determines how optimal solutions
are affected by changes within specified ranges.
– Changes in right-hand side (RHS) values.
– Changes in objective function coefficients.

• Managers must operate in dynamic environments with imprecise estimates of

coefficients.

30/39
Sensitivity Analysis
• Sensitivity analysis (or post-optimality analysis) determines how optimal solutions
are affected by changes within specified ranges.
– Changes in right-hand side (RHS) values.
– Changes in objective function coefficients.

• Managers must operate in dynamic environments with imprecise estimates of

coefficients.

• Sensitivity analysis is important for managers to ask ”what-if” questions about the
problem.

30/39
Graphical sensitivity Analysis
• We consider two cases:
1. Sensitivity of the optimum solution to changes in the availability of the resources
(right-hand side of the constraints)

2. Sensitivity of the optimum solution to changes in unit profit or unit cost (coefficients of the
objective function)

31/39
Changes in the Right-Hand side
• JOBCO manufactures two products on two machines.
• Processing times and revenues per unit are given as follows:
– Product 1: 2 hrs on machine 1, 1 hr on machine 2, $30 revenue per unit.

32/39
Changes in the Right-Hand side
• JOBCO manufactures two products on two machines.
• Processing times and revenues per unit are given as follows:
– Product 1: 2 hrs on machine 1, 1 hr on machine 2, $30 revenue per unit.
– Product 2: 1 hr on machine 1, 3 hrs on machine 2, $20 revenue per unit

32/39
Changes in the Right-Hand side
• JOBCO manufactures two products on two machines.
• Processing times and revenues per unit are given as follows:
– Product 1: 2 hrs on machine 1, 1 hr on machine 2, $30 revenue per unit.
– Product 2: 1 hr on machine 1, 3 hrs on machine 2, $20 revenue per unit

• Total daily processing time available for each machine is 8 hrs

• x1 and x2 represent the daily number of units of products 1 and 2.

32/39
Changes in the Right-Hand side
• JOBCO manufactures two products on two machines.
• Processing times and revenues per unit are given as follows:
– Product 1: 2 hrs on machine 1, 1 hr on machine 2, $30 revenue per unit.
– Product 2: 1 hr on machine 1, 3 hrs on machine 2, $20 revenue per unit

• Total daily processing time available for each machine is 8 hrs

• x1 and x2 represent the daily number of units of products 1 and 2.

Maximize z = 30 x1 + 20 x2
S.t 2 x1 + x2 ≤ 8
x1 + 3 x2 ≤ 8
x1 , x2 ≥ 0

32/39
 Changes in the Right-Hand side

• Increasing machine 1 capacity

from 8 to 9 hrs moves the
optimum solution to point G.
Rate of revenue change zG −zC
• Capacity change
= 9−8 .

$142−$128
• 9−8 = $14 \hour

• The point G should stays

between B and F.

• The dual price for machine 2

capacity is $2/hr.

33/39
Dual Prices
• The dual price is the rate of change of the objective function per unit change of a
resource.

• The abstract name ”dual” or ”shadow” price is standard in LP literature and

software packages.

• The dual price of $14/hr remains valid for changes in machine 1 capacity that move
its constraint parallel to itself to any point on the line segment BF.

• The dual price is only valid in the feasibility range (2.67 hr ≤ Machine 1 capacity ≤
16 hr), as calculated at points B and F.

• Changes outside this range produce a different dual price (worth per unit).

34/39
Changes in the RHS questions
• Question 1: If JOBCO can increase the capacity of both machines, which machine
should receive priority?

35/39
Changes in the RHS questions
• Question 1: If JOBCO can increase the capacity of both machines, which machine
should receive priority?

• Response: Priority should be given to machine 1, as each additional hour of

machine 1 increases revenue by $14, as opposed to only $2 for machine 2.

35/39
Changes in the RHS questions
• Question 1: If JOBCO can increase the capacity of both machines, which machine
should receive priority?

• Response: Priority should be given to machine 1, as each additional hour of

machine 1 increases revenue by $14, as opposed to only $2 for machine 2.

• Question 2: A suggestion is made to increase the capacities of machines 1 and 2 at

the additional cost of $10/hr for each machine. Is this advisable?

35/39
Changes in the RHS questions
• Question 1: If JOBCO can increase the capacity of both machines, which machine
should receive priority?

• Response: Priority should be given to machine 1, as each additional hour of

machine 1 increases revenue by $14, as opposed to only $2 for machine 2.

• Question 2: A suggestion is made to increase the capacities of machines 1 and 2 at

the additional cost of $10/hr for each machine. Is this advisable?

• Response: Only machine 1 should be considered for capacity increase, as the

additional net revenue per hour is 14-10=$4, compared to a net of 2-10=$-8 for
machine 2.

35/39
Changes in the RHS questions
• Question 3: If the capacity of machine 1 is increased from 8 to 13 hrs, how will this
increase impact the optimum revenue?

36/39
Changes in the RHS questions
• Question 3: If the capacity of machine 1 is increased from 8 to 13 hrs, how will this
increase impact the optimum revenue?

• Response: The proposed increase falls within the feasibility range for machine 1
and will result in a $14(13 - 8) =$70 increase in revenue, from $128 to $198 ( =$128
+ $70).

36/39
Changes in the RHS questions
• Question 3: If the capacity of machine 1 is increased from 8 to 13 hrs, how will this
increase impact the optimum revenue?

• Response: The proposed increase falls within the feasibility range for machine 1
and will result in a $14(13 - 8) =$70 increase in revenue, from $128 to $198 ( =$128
+ $70).

• Question 4: Suppose that the capacity of machine 1 is increased to 20 hrs, how will
this increase affect the optimum revenue?

36/39
Changes in the RHS questions
• Question 3: If the capacity of machine 1 is increased from 8 to 13 hrs, how will this
increase impact the optimum revenue?

• Response: The proposed increase falls within the feasibility range for machine 1
and will result in a $14(13 - 8) =$70 increase in revenue, from $128 to $198 ( =$128
+ $70).

• Question 4: Suppose that the capacity of machine 1 is increased to 20 hrs, how will
this increase affect the optimum revenue?

• Response: The proposed increase falls outside the feasibility range, and further
calculations are needed to determine the impact on optimum revenue.

36/39
Objective Coefficient Change

• Maximize z = c1 x1 + c2 x2 .
• How the optimum changes when we
change c1 and c2 .

37/39
Objective Coefficient Change

• Maximize z = c1 x1 + c2 x2 .
• How the optimum changes when we
change c1 and c2 .
• Changes in objective coefficients change
the slope of the isoprofit.

37/39
Objective Coefficient Change

• Maximize z = c1 x1 + c2 x2 .
• How the optimum changes when we
change c1 and c2 .
• Changes in objective coefficients change
the slope of the isoprofit.
• Optimum at C remains if objective
function is between BF and DE.

37/39
Objective Coefficient Change

• Maximize z = c1 x1 + c2 x2 .
• How the optimum changes when we
change c1 and c2 .
• Changes in objective coefficients change
the slope of the isoprofit.
• Optimum at C remains if objective
function is between BF and DE.
• Optimality range for coefficients
keeping optimum at C: 31 ≤ cc1 ≤ 21 .
2

37/39
Objective Coefficient Change questions
• Question 1: If unit revenues for Products 1 and 2 are changed to $35 and $25,
respectively, will the current optimum remain the same?
c1 35
• The solution at C will remain optimal because c2
= 25 = 1.4 remains within the
optimality range ( 13 , 2).

• Question 2: If the unit revenue of Product 2 is fixed at its current value c2 = $20,
what is the associated optimality range for the unit revenue for Product 1, c1 , that
will keep the optimum unchanged?

• The optimality range for c1 is: 20 × 1

3 ≤ c1 ≤ 2 × 20.

38/39
Conclusion
To sum up, we have covered the following key points:
• The two-phase method provides a viable approach for finding an initial feasible
solution for the Simplex method.

• While executing Simplex, one must consider its numerous special cases such as
degeneracy, unboundedness, and infeasibility to prevent potential issues.

• Cycling may occur in LP problems with degeneracy, which requires attention to

ensure convergence to an optimal solution.

• Graphical sensitivity analysis is a useful tool for investigating the impact of LP

parameter changes on the optimal solution, particularly in two dimensions.

39/39