Scribd
Upload
28 views28 pages

Lecture 11

The document discusses four special cases in engineering optimization: degeneracy, alternative optima, unbounded solutions, and infeasible solutions. Each case is explained with examples, highlighting how they affect the optimization process and potential issues that may arise, such as cycling and model misformulation. The document emphasizes the importance of correctly identifying and addressing these cases to ensure effective optimization.

Uploaded by

f20221605
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
28 views28 pages

Lecture 11

The document discusses four special cases in engineering optimization: degeneracy, alternative optima, unbounded solutions, and infeasible solutions. Each case is explained with examples, highlighting how they affect the optimization process and potential issues that may arise, such as cycling and model misformulation. The document emphasizes the importance of correctly identifying and addressing these cases to ensure effective optimization.

Uploaded by

f20221605
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 28

ENGINEERING OPTIMIZATION

ME F320/MF F320

BITS Pilani
Pilani Campus
3.5 Special Cases

2
Special Cases
Four special cases:
1. Degeneracy
2. Alternative optima
3. Unbounded solutions
4. Non-existing (or infeasible) solutions

3
1. Degeneracy
• A tie for minimum ratio may occur  broken
arbitrarily.
• At least one basic variable will be zero in next
iteration  new solution called “degenerate”.
• Degeneracy can cause iterations to cycle
indefinitely.
• Reveals possibility of at least one redundant
constraint.

4
1. Degeneracy: Example
Maximize
z = 3x1 + 9x2
subject to
x1 + 4x2 ≤ 8
x1 + 2x2 ≤ 4
x1, x2 ≥ 0

5
1. Degeneracy: Example

z column removed from tableau for convenience


6
1. Degeneracy: Example

• 3 lines pass through the optimum point (x1 = 0, x2 = 2)


• Point is overdetermined  one redundant constraint
 removed without changing the solution space.
No efficient computational techniques for identifying
redundant constraints! 
7
1. Degeneracy
• Can lead to cycling.
– In iterations 1 and 2, the objective does not improve
(z = 18)  possible for the simplex to never
complete  never improving the objective and
never satisfying optimality condition.
– Not very common.
– Algorithms for eliminating cycling are slow 
implemented when there is evidence of cycling.

8
1. Degeneracy
• Computer round-off error may actually create
degeneracy-like conditions.
• Iterations “stall” at a solution point  mimic
cycling.
• Commercial codes attempt to alleviate this
problem by periodically perturbing values of
basic variables (see Section 3.7).

9
2. Alternative optima
• LP problem may have an infinite number of
alternative optima when objective function is
parallel to a non-redundant binding
constraint (i.e., a constraint that is satisfied as
an equation at the optimal solution).

10
2. Alternative optima: Example
Maximize
z = 2x1 + 4x2
subject to
x1 + 2x2 ≤ 5
x1 + x2 ≤ 4
x1, x2 ≥ 0

11
2. Alternative optima: Example

12
2. Alternative optima: Example

z column removed from tableau for convenience

• Detection of alternative optimum  examine z-


equation coefficients of non-basic variables.
• Zero coefficient of x1  x1 can be made basic without
changing value of z  alternate solution is obtained.
13
2. Alternative optima: Example

z column removed from tableau for convenience

14
2. Alternative optima: Example

z column removed from tableau for convenience

• Deals with corner point optima only (points B and C).


15
2. Alternative optima
• All points (ොx1, xො 2) on line segment BC can be
written as ???

16
2. Alternative optima
• All points (𝑥ො 1, 𝑥ො 2) on line segment BC can be
written as:

• Can choose from many solutions without


deteriorating the objective.

17
3. Unbounded solution
• Variables may be increased indefinitely without
violating any of the constraints

• Model is poorly constructed.


– most likely irregularity is that some key constraints
have not been accounted for.
– estimates of constraint coefficients may be
inaccurate.

18
3. Unbounded solution: Example
Maximize
z = 2x1 + x2
subject to
x1 − x2 ≤ 10
2x1 ≤ 40
x1, x2 ≥ 0

Write the starting simplex table.

19
3. Unbounded solution: Example
Maximize
z = 2x1 + x2
subject to
x1 − x2 ≤ 10
2x1 ≤ 40
x1, x2 ≥ 0

z column removed from tableau for convenience


20
3. Unbounded solution: Example

z column removed from tableau for convenience

x2 can be increased indefinitely without violating any


constraints!

21
3. Unbounded solution: Example

22
4. Infeasible solution
• Doesn’t occur if all constraints are of type ≤
with nonnegative right-hand.
• Other types of constraints, penalized artificial
variables used to start the solution.
• If at least one artificial variable is positive in
optimum iteration  no feasible solution.
• Practically infeasible solution space means that
model is not formulated correctly.

23
4. Infeasible solution: Example
Maximize
z = 3x1 + 2x2
subject to
2x1 + x2 ≤ 2
3x1 + 4x2 ≥ 12
x1, x2 ≥ 0

Use the big M-method (M = 100) to start the


solution.
24
4. Infeasible solution: Example

z column removed from tableau for convenience

25
4. Infeasible solution: Example

• R is positive (= 4)  LPP is infeasible.

26
4. Infeasible solution: Example

27
Thank you!

28

You might also like