Module 1
Linear Programming Problems
Part 5 – Variants of Simplex Method
Course Instructor: Dr. Sahaya Jenifer A
Special cases of Simplex Method
(i) Infeasible solution
(ii) Unbounded solution
(iii) Multiple optimal solution
(iv) Unrestricted variables
(iv) Degeneracy (Tie in leaving variable)
Infeasible solution:
- If the 𝑍 −row values satisfy the
optimality conditions, but atleast one artificial
variable is still in the basis with positive RHS,
then the problem has no feasible solution.
- It reverses the constraint inequality.
- This is called pseudo-optimal solution.
Example:
Max 𝑍 = 3𝑥 + 2𝑥
subject to
2𝑥 + 𝑥 ≤ 2
3𝑥 + 4𝑥 ≥ 12
𝑥 ,𝑥 ≥ 0
Standard form:
Initial Basic Feasible Solution:
Rewrite objective function:
Iteration 1:
Basic 𝑥 𝑥 𝑠 𝑠 𝑅 RHS
Optimal condition is satisfied?
Iteration 2:
Basic 𝑥 𝑥 𝑠 𝑠 𝑅 RHS
Optimal condition is satisfied?
Unbounded solution:
- If all the entering column coefficients
are ≤ 0, then the LPP has an unbounded
solution.
- Increasing the value of the entering
variable improves the objective function
without limit.
Example:
Max 𝑍 = 2𝑥 + 𝑥
subject to
𝑥 − 𝑥 ≤ 10
2𝑥 − 𝑥 ≤ 40
𝑥 ,𝑥 ≥ 0
Standard form:
Initial basic feasible solution:
Iteration 1:
Basic 𝑥 𝑥 𝑠 𝑠 RHS
𝑠
𝑠
𝑍
Optimal condition is satisfied?
Iteration 2:
Basic 𝑥 𝑥 𝑠 𝑠 RHS
Optimal condition is satisfied?
Iteration 3:
Basic 𝑥 𝑥 𝑠 𝑠 RHS
Optimal condition is satisfied?
Alternate (or) Multiple Optimal Solution
- If the non-basic variable has zero coefficient
in the 𝑍 row at the optimum level, then there
exists an alternate optimal solution for the given
LPP.
- That non-basic variable entering into the
basis does not change the objective function
value, so the new solution is also optimal.
Example:
Max 𝑍 = 4𝑥 + 10𝑥
subject to
2𝑥 + 𝑥 ≤ 10
2𝑥 + 5𝑥 ≤ 20
2𝑥 + 3𝑥 ≤ 18
𝑥 ,𝑥 ≥ 0
Standard form:
Initial basic feasible solution:
Iteration 1:
Basic 𝑥 𝑥 𝑠 𝑠 𝑠 RHS
𝑠
𝑠
𝑠
𝑍
Optimal condition is satisfied?
Iteration 2:
Basic 𝑥 𝑥 𝑠 𝑠 𝑠 RHS
𝑍
Optimal condition is satisfied?
Iteration 3:
Basic 𝑥 𝑥 𝑠 𝑠 𝑠 RHS
𝑍
Optimal condition is satisfied?
Tie in the choice of entering variable
- when more than one variable have the
most negative in the 𝑍 −row, any one of
them can be selected arbitrarily as the
entering variable. But the wrong choice may
increase the no. of iterations to reach the
optimum.
- Give preference to decision variables
over slack variables in case of a tie.
Tie in the choice of leaving variable
(degeneracy)
- when more than one variable have the
minimum ratio, atleast one of the basic
variable will be zero in the next iteration →
degenerate solution.
- leads to possible cycling or slower
progress.
-To avoid this, if any artificial variable occurs
in the tied rows, choose it as the leaving
variable.
- Otherwise, divide each tied row by the pivot
column element and compare starting from
the left. Choose the variable corresponding to
the min value as the leaving variable.
Example:
Max 𝑍 = 𝑥 + 2𝑥 + 𝑥
subject to
2𝑥 + 𝑥 − 𝑥 ≤ 2
−2𝑥 + 𝑥 − 5𝑥 ≥ −6
4𝑥 + 𝑥 + 𝑥 ≤ 6
𝑥 ,𝑥 ,𝑥 ≥ 0
Standard form:
Initial basic feasible solution:
Iteration 1:
Basic 𝑥 𝑥 𝑥 𝑠 𝑠 𝑠 RHS
𝑠
𝑠
𝑠
𝑍
Optimal condition is satisfied?
Iteration 2:
Basic 𝑥 𝑥 𝑥 𝑠 𝑠 𝑠 RHS
𝑍
Optimal condition is satisfied?
Iteration 3:
Basic 𝑥 𝑥 𝑥 𝑠 𝑠 𝑠 RHS
𝑍
Optimal condition is satisfied?
Unrestricted variables
If any variable is unrestricted, it can be
expresses as the difference between two non-
negative variables.
If 𝑥 is unrestricted →
𝑥 =𝑥 −𝑥 , 𝑥 ,𝑥 ≥ 0
Example:
Max 𝑍 = 8𝑥
subject to
𝑥 −𝑥 ≥0
2𝑥 + 3𝑥 ≤ −6
𝑥 , 𝑥 are unrestricted.
Standard form:
Initial Basic Feasible Solution:
Rewrite objective function:
Iteration 1:
Basic 𝑥 ′ 𝑥 ′′ 𝑥 𝑥 𝑠 𝑠 𝑅 RHS
𝑍
Optimal condition is satisfied?
Iteration 2:
Basic 𝑥 ′ 𝑥 ′′ 𝑥 𝑥 𝑠 𝑠 RHS
𝑍
Optimal condition is satisfied?
Iteration 3:
Basic 𝑥 ′ 𝑥 ′′ 𝑥 𝑥 𝑠 𝑠 RHS
𝑍
Optimal condition is satisfied?
DIY:
Max 𝑍 = 3𝑥 + 9𝑥
subject to
𝑥 + 4𝑥 ≤ 8
𝑥 + 2𝑥 ≤ 4
𝑥 ,𝑥 ≥ 0
DIY:
Max 𝑧 = 6𝑥 + 4𝑥
subject to
2𝑥 + 3𝑥 ≤ 30
3𝑥 + 2𝑥 ≤ 24
𝑥 +𝑥 ≥3
𝑥 ,𝑥 ≥ 0
DIY:
Max 𝑍 = 2𝑥 + 3𝑥
subject to
𝑥 − 2𝑥 ≤ 0
−2𝑥 + 3𝑥 ≤ −6
𝑥 , 𝑥 are unrestricted.
DIY:
Max 𝑍 = 2𝑥 + 𝑥
subject to
3/2 𝑥 + 𝑥 ≤ 6
𝑥 ≤2
𝑥 +𝑥 ≥7
−𝑥 + 𝑥 ≥ 4
𝑥 ,𝑥 ≥ 0
DIY:
Max 𝑍 = 2𝑥 + 𝑥 + 𝑥
subject to
4𝑥 + 6𝑥 + 3𝑥 ≤ 8
3𝑥 − 6𝑥 − 4𝑥 ≤ 1
2𝑥 + 3𝑥 − 5𝑥 ≥ 4
𝑥 ,𝑥 ,𝑥 ≥ 0
