CONSTRAINED
OPTIMIZATION WITH
INEQUALITY
CONSTRAINTS
Dr. Purba Roy Choudhury
Email: purba.roychoudhury@thebges.edu.in
B.Sc. Economics (Honours)
Semester II, Paper CC-2-4:Mathematical Methods in Economics II
Unit 2: Multi-Variable Optimization
Constrained Optimization with Inequality Constraints: Kuhn Tucker Conditions
Kuhn Tucker Conditions
• In the Classical optimization problem, with no explicit restrictions on
the signs of the choice variables, and with no inequalities in the
constraints, the first order conditions for a relative or local extremum
is simply the first order partial derivatives of the (smooth) Lagrangian
function with respect to all the choice variables and the Lagrange
multiplier will be zero.
• In non-linear programming, there exists a similar type of first order
condition, known as Kuhn Tucker conditions.
• While the Classical first order conditions are always necessary, the
Kuhn Tucker conditions can not be accorded the status of necessary
conditions unless a certain proviso is satisfied.
• Under certain specific circumstances, the Kuhn Tucker conditions turn
out to be sufficient conditions or even necessary-and-sufficient
conditions as well.
• The Kuhn Tucker conditions are the single most important analytical
result in non-linear programming and its understanding involves two
steps.
Step1: Effect of Non-negativity restrictions
• Consider a problem with non-negativity restrictions and no other constraint
where the function f is assumed to be differentiable.
• With the restriction 𝑥1 ≥ 0, there are three possibilities
Case I: Interior Solution:
If the local maximum occurs at point A in the region of the
shaded feasible region in the diagram (a).
Here
Step1: Effect of Non-negativity restrictions
Case II : Boundary Solution:
The local maximum occur on the vertical axis where
𝑥1 = 0 at a point B on diagram (b)
Here the first order condition is valid
Here
Case III : Boundary Solution:
The points C and D in diagram (c)is a local maximum
provided the candidate point is higher than the
neighbouring points in the feasible region
Here the maximum point is characterized not only by
but also by the inequality
Here
Note: the inequality is ruled out for a point where the curve is
upward rising. There cannot be a maximum even if the point is located on
the vertical axis, such as point E in diagram (a)
Step1: Effect of Non-negativity restrictions
Step1: Effect of Non-negativity restrictions
Step1: Effect of Non-negativity restrictions
Step 2:Effect of Inequality Constraints
Step 2:Effect of Inequality Constraints
Step 2:Effect of Inequality Constraints
Step 2:Effect of Inequality Constraints
Step 2:Effect of Inequality Constraints
Step 2:Effect of Inequality Constraints
Step 2:Effect of Inequality Constraints
Example
Example
Example
Interpreting the Lagrange Multiplier
Interpreting the Lagrange Multiplier
Interpreting the Lagrange Multiplier
Case of n variables and m constraints
Minimization
Minimization
Kuhn Tucker Conditions
• Maximization
• Minimization
Example
Example
Applications: War time Rationing
Applications: War time Rationing
Applications: War time Rationing
Applications: War time Rationing
Applications: War time Rationing
Applications: War time Rationing
Applications: War time Rationing
Applications: Peak Load Pricing
Applications: Peak Load Pricing
Applications: Peak Load Pricing
Applications: Peak Load Pricing
Applications: Peak Load Pricing
Applications: Peak Load Pricing
Applications: Peak Load Pricing
Applications: Peak Load Pricing
Applications: Peak Load Pricing
References:
• Alpha C. Chiang and Kavin Wainwright: Fundamental
Methods of Mathematical Economics
• K. Sydsaeter and P. Hammond: Mathematics for
Economic Analysis
• Carl Simon and Lawrence Blume: Mathematics for
Economists
• Archibald, G.C. and Lipsey, R.G.: An Introduction to
Mathematical Treatment of Economics
• Henderson, J.M. and Quandt, R.E.: Microeconomic
Theory : A Mathematical Approach
