Operations Research

LECTURE 6
Dual Problem in
Linear Programming
The term ‘dual’, in general, implies two or double. The concept of duality is very useful in
mathematics, physics, statistics, engineering, and managerial decision-making.

For example, in a two-person game theory, one competitor’s problem is the dual of the
opponent’s problem.

In linear programming, duality implies that each linear programming problem can be analyzed in
two different ways but would have equivalent solutions. Any LP problem (either maximization or
minimization) can be stated in another equivalent form based on the same data.

The new LP problem is called a dual linear programming problem or in short dual. In general, it is
immaterial which of the two problems is called primal or dual, since the dual of the dual is primal.

For example, consider the problem of production planning. The production manager attempts to
determine quantities for each product to be produced with an objective to optimize the use of
available resources so that profit is maximum.

But through a dual LP problem approach, he may develop a production plan that optimizes resource
utilization so that the marginal opportunity cost of each unit of a resource is equal to its marginal
return (also called shadow price).

The shadow price indicates an additional price to be paid to obtain one additional unit of the
resources in order to maximize profit under the resource constraints. If a resource is not completely
used, i.e. there is slack, then its marginal return is zero.

𝐶ℎ𝑎𝑛𝑔𝑒 𝑖𝑛 𝑜𝑝𝑡𝑖𝑚𝑎𝑙 𝑜𝑏𝑗𝑒𝑐𝑡𝑖𝑣𝑒 𝑓𝑢𝑛𝑐𝑡𝑖𝑜𝑛 𝑣𝑎𝑙𝑢𝑒

𝑆ℎ𝑎𝑑𝑜𝑤 𝑃𝑟𝑖𝑐𝑒 =
𝑈𝑛𝑖𝑡 𝑐ℎ𝑎𝑛𝑔𝑒 𝑖𝑛 𝑡ℎ𝑒 𝑎𝑣𝑎𝑖𝑙𝑎𝑏𝑖𝑙𝑖𝑡𝑦 𝑜𝑓 𝑟𝑒𝑠𝑜𝑢𝑟𝑐𝑒

Formulation of Dual Linear Programming Problem

There are two important forms of primal and dual LP problems, namely the symmetrical (canonical)
form and the standard form.

Symmetrical Form
Suppose the primal LP problem is given in the form
𝑀𝑎𝑥𝑖𝑚𝑖𝑧𝑒 𝑍! = 𝑐" 𝑥" + 𝑐# 𝑥# + . . . + 𝑐$ 𝑥$
subject to the constraints
𝑎"" 𝑥" + 𝑎"# 𝑥# + . . . + 𝑎"$ 𝑥$ ≤ 𝑏"
𝑎#" 𝑥" + 𝑎## 𝑥# + . . . + 𝑎#$ 𝑥$ ≤ 𝑏#
.
.
.
𝑎%" 𝑥" + 𝑎%# 𝑥# + . . . + 𝑎%$ 𝑥$ ≤ 𝑏%
And 𝑥" , 𝑥# . . . , 𝑥$ ≥ 0

2
Then the corresponding dual LP problem is written as
𝑴𝒊𝒏𝒊𝒎𝒊𝒛𝒆 𝒁𝒚 = 𝒃𝟏 𝒚𝟏 + 𝒃𝟐 𝒚𝟐 + . . . + 𝒃𝒎 𝒚𝒎
subject to the constraints
𝑎"" 𝑦" + 𝑎"# 𝑦# + . . . + 𝑎%" 𝑦% ≤ 𝑐"
𝑎#" 𝑦" + 𝑎## 𝑦# + . . . + 𝑎%# 𝑦% ≤ 𝑐#
.
.
.
𝑎"$ 𝑦" + 𝑎#$ 𝑦# + . . . + 𝑎%$ 𝑦$ ≤ 𝑐%
𝑦" , 𝑦# . . . , 𝑦% ≥ 0

Interpreting Dual Variables in LPP

Interpreting dual variables in linear programming problems (LPP) is crucial for understanding the
economic significance of the constraints and their impact on the objective function. Dual variables,
also known as shadow prices or Lagrange multipliers, provide insights into the rate of change of the
objective function with respect to changes in the constraint coefficients.

Here's how to interpret dual variables in LPP:

1. Economic Interpretation:
Dual variables represent the marginal value or opportunity cost associated with relaxing or
tightening a constraint by one unit while keeping the rest of the problem unchanged. In other
words, they indicate how much the optimal objective value of the problem would change if the
constraint's right-hand side (or coefficient) were modified slightly.

2. Shadow Prices:
For each constraint in the primal problem, there is a corresponding dual variable. The dual variable
associated with a constraint is often referred to as its ‘shadow price.’ The shadow price represents
the change in the optimal objective function value per unit increase on the right-hand side of the
corresponding constraint, assuming all other variables remain optimal.

3. Interpretation for Non-Negative Constraints:

For a constraint that has a non-negative (≥) sign in the primal problem, the associated dual variable
indicates the increase in the objective function value for each unit increase in the right-hand side of
the constraint.

3
4. Interpretation for Non-Positive Constraints:
For a constraint that has a non-positive (≤) sign in the primal problem, the associated dual variable
indicates the decrease in the objective function value for each unit increase in the right-hand side
of the constraint.

5. Zero Dual Variable:

If a dual variable is zero, it indicates that the corresponding constraint is not binding (not active) at
the optimal solution. This means that a small change in the right-hand side of that constraint will
not affect the optimal solution or the objective function value.

6. Range of Validity:
The dual variable is only valid within certain ranges of constraint changes. If the constraint is
already fully binding (active), further relaxing it won't change the objective value, and the
corresponding dual variable might not hold an economic interpretation.

In general, the primal-dual relationship between a pair of LP problems can be expressed

as follows:

Relationship Between Primal and Dual Problem

If Primal Then Dual

Objective is to maximize Objective is to minimize

Jth primal variable, 𝑥* Jth dual constraint

ith primal variable ith dual constraint, 𝑦+

Primal variable 𝑥* unrestricted in sign Dual constraint j is =type

Primal constraint i =type Dual variable 𝑦+ is unrestricted in sign

Dual constraints ≤type Dual Constraint ≥ type

4
Solving Linear Programming Problems using Dual Problem
Question 1: Write the dual to the following LP problem.

Maximize 𝑍 = 𝑥" – 𝑥# + 3𝑥,

subjects to the constraints
i. 𝑥" + 𝑥# + 𝑥, ≤ 10
ii. 2𝑥" − 𝑥# −𝑥, ≤ 2
ii. 2𝑥" − 2𝑥# −3𝑥, ≤ 6,
and
𝑥" , 𝑥# , 𝑥, ≥ 0

Solution:
In the given LP problem there are 𝑚 = 3 constraints and 𝑛 = 3 variables. Thus, there must be 𝑚 =
3 dual variables and n = 3 constraints. Further, the coefficients of the primal variables,𝑐" = 1, 𝑐# =
– 1, 𝑐, = 3 become right-hand side constants of the dual.

The right-hand side constants 𝑏" = 10, 𝑏# = 2, 𝑏, = 6 become the coefficients in the dual objective
function. Finally, the dual must have a minimizing objective function with all ≥ type constraints.

If 𝑦" , 𝑦# 𝑎𝑛𝑑, 𝑦, are dual variables corresponding to three primal constraints in the given order, the
resultant dual is

Minimize 𝑍- = 10𝑦" + 2𝑦# + 6𝑦,

Subjects to the constraints
i. 𝑦" + 2𝑦# + 2𝑦, ≥ 1,
ii. 𝑦" − 𝑦# − 2𝑦, ≥ −1
iii. 𝑦" − 𝑦# − 3𝑦, ≥ 3
iv. 𝑦" , 𝑦# , 𝑦, ≥ 0

Question 2: Write the dual to the following LP problem.

Minimize 𝑍 = 3𝑥" – 2𝑥# + 4𝑥,
subjects to the constraints
i. 3𝑥" + 5𝑥# + 4𝑥, ≥ 7,
ii. 6𝑥" + 𝑥# +3𝑥, ≥ 4,
iii. 7𝑥" − 2𝑥# −𝑥, ≤ 10,
iv. 𝑥" − 2𝑥# + 5𝑥, ≥ 3,
v. 4𝑥" + 7𝑥# − 2𝑥, ≥ 2,

5
and
𝑥" , 𝑥# , 𝑥, ≥ 0

Solution:
Since the objective function of the given LP problem is of depreciation, the direction of each
inequality has to be changed to ≥ type by multiplying both sides by – 1. The standard primal LP
problem so obtained is

Minimize 𝒁𝒙 = 𝟑𝒙𝟏 − 𝟐𝒙𝟐 + 𝟒𝒙𝟑

Subject to the constraints

i. 3𝑥" + 5𝑥# + 4𝑥, ≥ 7

ii. 6𝑥" + 𝑥# + 3𝑥, ≥ 4
iii. −7𝑥" + 2𝑥# + 𝑥, ≥ −10
iv. 𝑥" − 2𝑥# + 5𝑥, ≥ 3x1
v. 4𝑥" + 7𝑥# − 2𝑥, ≥ 2

and
If 𝑦" , 𝑦# , 𝑦, , 𝑦0 and 𝑦1 are dual variables corresponding to the five primal constraints in the given
order, the dual of this primal LP problem is stated as
Maximize Zy = 7𝑦" + 4𝑦# – 10𝑦, + 3𝑦0 + 2𝑦1
subject to the constraints
i. 3𝑦" + 6𝑦# – 7𝑦, + 𝑦0 + 4𝑦1 ≤ 3,
ii. 5𝑦" + 𝑦# + 2𝑦, – 2𝑦0 + 7𝑦1 ≤ –2
iii. 4𝑦" + 3𝑦# + 𝑦, + 5𝑦0 – 2𝑦1 ≤ 4

and 𝑦" , 𝑦# , 𝑦, , 𝑦, , 𝑦0 ≥ 0

Practice Questions
1. Max 𝒁𝒙 = 𝟐𝒙𝟏 + 𝟓𝒙𝟐 + 𝟔𝒙𝟑 subjects to
i. 5𝑥" + 6𝑥# − 𝑥, ≤ 3
ii. −2𝑥" + 3𝑥# + 4𝑥, ≤ 4
iii. 𝑥" − 5𝑥# + 3𝑥, ≤ 1
iv. −3𝑥" − 3𝑥# + 7𝑥, ≤ 6
and 𝑥" , 𝑥# , 𝑥, ≥ 0

6
2. Min 𝒁𝒙 = 𝟕𝒙𝟏 + 𝟑𝒙𝟐 + 𝟖𝒙𝟑 subjects to

i. 8𝑥" + 2𝑥# + 𝑥, ≥ 3
ii. 3𝑥" + 6𝑥# + 4𝑥, ≥ 4
iii. 4𝑥" + 𝑥# + 5𝑥, ≥ 1
iv. 𝑥" + 5𝑥# + 2𝑥, ≥ 7

and 𝑥" , 𝑥# , 𝑥, ≥ 0

References:

§ Hillier, F. S., & Lieberman, G. J. (2013). Introduction to operations research. New York, NY: McGraw-Hill.
§ Winston, W. L. (2014). Operations research: Applications and algorithms. Boston, MA: Cengage Learning.
§ Taha, H. A. (2016). Operations research: An introduction. Upper Saddle River, NJ: Pearson.