➢ The role of duaty theory insensitivity analysis
Duality theory plays a significant role in sensitivity analysis, particularly within the
context of linear programming. Here's how:
1. Providing Shadow Prices
• Definition: Shadow prices (also known as dual prices or marginal values) represent the
change in the optimal objective function value of the primal problem per unit increase in
the right-hand side of a constraint. Shadow prices are directly obtained from the optimal
solution of the dual problem.
• Use in Sensitivity Analysis: Shadow prices help assess the impact of changes in resource
availability or constraint requirements. For example, if a constraint represents the
availability of labor hours, the shadow price tells you how much the profit would increase
if you had one more labor hour available. This information is crucial for resource
allocation decisions.
2. Determining Allowable Changes
Duality theory helps determine the range within which the objective function coefficients
can change without affecting the optimality of the current solution (i.e., the same
variables remain in the optimal basis). Similarly, it helps determine the range within
which the right-hand side values of the constraints can change without making the current
solution infeasible (i.e., the same constraints remain binding).
• Use in Sensitivity Analysis: These ranges provide valuable insights into the robustness
of the optimal solution. If a coefficient or a right-hand side value is likely to change
within its allowable range, you can be confident that your current solution will still be
optimal or feasible.
Abera.A(Ph.D) Page 1
3. Economic Interpretation
• Dual Problem as an Economic Model: The dual problem can often be interpreted as an
economic model in its own right. For example, if the primal problem is a production
problem that seeks to maximize profit, the dual problem can be interpreted as a resource
valuation problem that seeks to minimize the cost of resources.
• Use in Sensitivity Analysis: This economic interpretation can provide additional insights
into the sensitivity of the optimal solution to changes in economic parameters.
4. Computational Efficiency
• Solving the Dual Instead of the Primal: In some cases, it may be computationally more
efficient to solve the dual problem instead of the primal problem, especially if the dual
has fewer constraints or variables.
• Use in Sensitivity Analysis: If you solve the dual to find the optimal solution, you
automatically obtain the shadow prices, which are essential for sensitivity analysis.
Example
Consider a company that produces two products, A and B, using two resources, labor and
materials. The primal problem might be to maximize profit subject to constraints on labor
and materials availability.
• The shadow price of the labor constraint (obtained from the dual solution) would tell you
how much the profit would increase if the company had one more unit of labor available.
• The range of feasibility for the materials constraint would tell you how much the
availability of materials could change without affecting the feasibility of the current
production plan.
In summary, duality theory provides powerful tools for sensitivity analysis in linear
programming by:
Providing shadow prices for assessing the impact of changes in constraints.
Determining allowable changes in objective function coefficients and constraint
right-hand sides.
Offering economic interpretations of the problem.
Potentially improving computational efficiency.
Abera.A(Ph.D) Page 2
By understanding and utilizing duality theory, you can gain a deeper
understanding of the sensitivity of your linear programming models and make
more informed decisions.
➢ Essence of sensitivity analysis & applying sensitivity analysis
Sensitivity analysis is a way to predict the impact of changing variables on the outcome
of a decision or a project. Here's a breakdown of its essence and how to apply it:
Essence of Sensitivity Analysis
At its core, sensitivity analysis is about understanding "what-if" scenarios. It helps
answer questions like:
• "What if our sales are lower than expected?"
• "What if the cost of raw materials increases?"
• "How will our profit be affected if we delay the project by a month?"
By examining these scenarios, you can:
❖ Identify critical factors: Pinpoint which variables have the biggest impact on
your outcome.
❖ Assess risk: Understand the potential range of outcomes and the likelihood of
undesirable results.
❖ Make informed decisions: Choose the best course of action based on a clear
understanding of the potential consequences.
❖ Communicate effectively: Explain the risks and uncertainties to stakeholders in a
clear and concise way.
Applying Sensitivity Analysis
Here's a step-by-step guide to applying sensitivity analysis:
1. Identify the key outcome: Determine the metric you want to analyze (e.g., profit, net
present value, project completion time).
2. Identify the input variables: List the factors that could affect the outcome (e.g., sales
volume, costs, interest rates, project duration).
3. Establish the base case: Define the most likely or expected values for each input
variable. This will serve as your benchmark.
Abera.A(Ph.D) Page 3
4. Define the scenarios: Determine the range of possible values for each input variable.
You can use:
o One-at-a-time analysis: Change one variable at a time while holding others constant.
o Scenario analysis: Create a few distinct scenarios with different combinations of input
values (e.g., best-case, worst-case, most likely).
o Simulation: Use statistical techniques (like Monte Carlo simulation) to generate a large
number of random scenarios.
5. Calculate the outcome for each scenario: Use your model or calculations to determine
the outcome for each scenario.
6. Analyze the results: Compare the outcomes across different scenarios to see how
sensitive the outcome is to changes in each input variable.
Example
Let's say you're analyzing the profitability of a new product:
1. Key outcome: Profit
2. Input variables:
o Selling price
o Variable cost per unit
o Fixed costs
o Sales volume
3. Base case:
o Selling price: $100
o Variable cost: $60
o Fixed costs: $50,000
o Sales volume: 1,000 units
4. Scenarios:
o Increase selling price by 10%
o Increase variable cost by 5%
o Decrease sales volume by 20%
5. Calculate outcomes: Calculate the profit for each scenario.
Abera.A(Ph.D) Page 4
6. Analyze results: Compare the profit in each scenario to the base case to see how
sensitive the profit is to changes in each variable.
Tools for Sensitivity Analysis
• Spreadsheet software: Tools like Microsoft Excel or Google Sheets are commonly used
for basic sensitivity analysis.
• Specialized software: There are dedicated software packages for more complex
simulations and risk analysis.
By understanding the essence of sensitivity analysis and applying it effectively, you can
make better decisions, manage risks more effectively, and improve your chances of
success.
➢ Change in objective function coefficient
A change in the objective function coefficient refers to a modification in the value
associated with a decision variable in the objective function of a linear programming (LP)
problem. This change can affect the optimal solution of the LP. Sensitivity analysis helps
us understand how much these coefficients can change without altering the optimal
solution.
Impact of Change in Objective Function Coefficient
1. Change in Optimal Objective Value: Even if the optimal solution (the values of the
decision variables) remains the same, a change in the objective function coefficient will
directly change the optimal objective value (e.g., the maximum profit or minimum cost).
2. Change in Optimal Solution: If the change in the coefficient is large enough, it can
cause the optimal solution itself to change. This means a different set of decision
variables will become optimal.
Range of Optimality
✓ This range determines how much an objective function coefficient can increase or
decrease without changing the optimal solution (the values of the decision variables).
✓ Within this range, only the optimal objective value will change.
✓ If the change exceeds this range, the current optimal solution will no longer be optimal,
and a new solution will be found.
Determining the Range of Optimality
The range of optimality can be determined using various methods, including:
Abera.A(Ph.D) Page 5
1. Graphical Method: For LP problems with two decision variables, the graphical method
can be used to visualize the range of optimality. By changing the slope of the objective
function line, you can observe how much it can vary before it intersects a different corner
point of the feasible region (which would represent a new optimal solution).
2. Simplex Method and Sensitivity Reports: The simplex method, an algorithm for
solving LP problems, provides sensitivity reports as part of its output. These reports
include the allowable increase and allowable decrease for each objective function
coefficient.
3. Dual Problem: Duality theory provides insights into the range of optimality. Changes in
objective function coefficients in the primal problem correspond to changes in the
constraints of the dual problem.
Example
Consider the following LP problem:
Maximize Z = 3x₁ + 2x₂
Subject to:
• x₁ + x₂ ≤ 10
• 2x₁ + x₂ ≤ 16
• x₁, x₂ ≥ 0
Suppose the optimal solution is x₁ = 6, x₂ = 4, and Z = 26.
Practical Implications
• Robustness of the Solution: The range of optimality indicates how robust the optimal
solution is to changes in the estimates of the objective function coefficients. If the range
is wide, the solution is less sensitive to estimation errors.
• Decision Making: This information is crucial for decision-making. If you anticipate that
an objective function coefficient might change, you can assess whether this change will
affect your current optimal strategy.
• Negotiation: In some situations, objective function coefficients might be subject to
negotiation. Sensitivity analysis can help you understand the impact of different
negotiation outcomes on your optimal solution.
By analyzing the change in objective function coefficients, you can gain a deeper
understanding of the behavior of your LP model and make more informed decisions.
Abera.A(Ph.D) Page 6
➢ A change in a parameter of the constraints
A change in a parameter of the constraints in a linear programming (LP) problem refers
to a modification in the values that define the constraints. These parameters can be:
• Right-hand side (RHS) values: These represent the limits or availability of resources.
For example, in a constraint like "2x₁ + 3x₂ ≤ 50," the 50 is the RHS value, representing
the maximum available units of a resource.
• Constraint coefficients: These are the coefficients of the decision variables in the
constraints. For example, in the same constraint, the 2 and 3 are the constraint
coefficients.
Impact of Change in Constraint Parameters
Changes in constraint parameters can have several effects on the LP problem:
1. Change in Feasible Region: The feasible region is the set of all possible solutions that
satisfy all constraints. Changing constraint parameters can expand, shrink, or shift the
feasible region.
2. Change in Optimal Solution: A change in the feasible region can lead to a change in the
optimal solution. The optimal solution is always located at a corner point of the feasible
region, and changing the region can change which corner point is optimal.
3. Change in Optimal Objective Value: Even if the optimal solution remains at the same
corner point, a change in the constraints can still affect the optimal objective value.
Sensitivity Analysis for Constraint Parameters
Sensitivity analysis helps us understand how much these parameters can change without
drastically altering the optimal solution or its value. Key concepts include:
1. Range of Feasibility:
o This range determines how much the RHS value of a constraint can increase or decrease
without making the current optimal solution infeasible (i.e., it still satisfies all
constraints).
o Within this range, the optimal solution might change, but it will still be within the defined
constraints.
o If the change exceeds this range, the current optimal solution will become infeasible, and
a new optimal solution will be found.
2. Shadow Prices (Dual Prices):
Abera.A(Ph.D) Page 7
o Shadow prices, obtained from the dual problem, indicate the change in the optimal
objective value for a one-unit increase in the RHS of a constraint.
o They provide valuable information about the marginal value of resources.
o A shadow price of zero indicates that the constraint is non-binding (there is slack or
surplus), meaning a small change in the RHS will not affect the optimal objective value.
Example
Consider the following LP problem:
Maximize Z = 3x₁ + 2x₂
Subject to:
• x₁ + x₂ ≤ 10 (Resource A)
• 2x₁ + x₂ ≤ 16 (Resource B)
• x₁, x₂ ≥ 0
Suppose the optimal solution is x₁ = 6, x₂ = 4, and Z = 26.
• If sensitivity analysis shows that the range of feasibility for Resource A (RHS = 10) is
from 8 to 12, it means that the availability of Resource A can be between 8 and 12 units
without making the current optimal solution infeasible.
• If the shadow price for Resource B is 1, it means that if we increase the availability of
Resource B by one unit (from 16 to 17), the optimal objective value will increase by 1
(from 26 to 27).
Practical Implications
• Resource Management: The range of feasibility and shadow prices help in making
decisions about resource allocation. For example, if a resource has a high shadow price, it
might be worthwhile to acquire more of that resource.
• Contingency Planning: Sensitivity analysis helps in understanding the impact of
potential changes in resource availability or constraint requirements, allowing for better
contingency planning.
• Negotiation: In situations where resource availability is subject to negotiation, shadow
prices can provide valuable information for determining fair prices.
By analyzing the change in constraint parameters, you can gain a deeper understanding of
the robustness of your LP model and make more informed decisions about resource
allocation and other related aspects.
Abera.A(Ph.D) Page 8