Nonlinear Programming

Chapter 10

Copyright © 2010 Pearson Education, Inc. Publishing as Prentice Hall 10-1

Chapter Topics

■ Nonlinear Profit Analysis

■ Constrained Optimization

■ Solution of Nonlinear Programming Problems with Excel

■ Nonlinear Programming Model with Multiple Constraints

■ Nonlinear Model Examples

10-2
Overview

■ Problems that fit the general linear programming format but

contain nonlinear functions are termed nonlinear programming
(NLP) problems.
■ Solution methods are more complex than linear programming
methods.
■ Determining an optimal solution is often difficult, if not
impossible.
■ Solution techniques generally involve searching a solution surface
for high or low points requiring the use of advanced mathematics.

10-3
Optimal Value of a Single Nonlinear Function
Basic Model
Profit function, Z, with volume
independent of price:
Z = vp - cf - vcv
where v = sales volume
p = price
cf = unit fixed cost
cv = unit variable cost

Add volume/price relationship:

v = 1,500 - 24.6p

Figure 10.1 Linear Relationship of Volume to Price

10-4
Optimal Value of a Single Nonlinear Function

With fixed cost (cf = $10,000) and variable cost (cv = $8):
Profit, Z = 1,696.8p - 24.6p2 - 22,000

Figure 10.2 The Nonlinear Profit Function

10-5
Optimal Value of a Single Nonlinear Function
Maximum Point on a Curve
■ The slope of a curve at any point is equal to the derivative of the
curve’s function.
■ The slope of a curve at its highest point equals zero.

Figure 10.3 Maximum profit for the profit function

10-6
Optimal Value of a Single Nonlinear Function
Solution Using Calculus
Z = 1,696.8p - 24.6p2 - 2,000
dZ/dp = 1,696.8 - 49.2p
=0
p = 1696.8/49.2
= $34.49
v = 1,500 - 24.6p
v = 651.6 pairs of jeans
Z = $7,259.45

Figure 10.4
10-7
Constrained Optimization in Nonlinear Problems
Definition

■ A nonlinear problem containing one or more constraints becomes a

constrained optimization model or a nonlinear programming
(NLP) model.
■ A nonlinear programming model has the same general form as the
linear programming model except that the objective function and/or
the constraint(s) are nonlinear.
■ Solution procedures are much more complex and no guaranteed
procedure exists for all NLP models.

10-8
Constrained Optimization in Nonlinear Problems
Graphical Interpretation (1 of 3)
Effect of adding constraints to nonlinear problem:

Figure 10.5 Nonlinear Profit Curve for the Profit Analysis Model
10-9
Constrained Optimization in Nonlinear Problems
Graphical Interpretation (2 of 3)

Figure 10.6 A Constrained Optimization Model

10-10
Constrained Optimization in Nonlinear Problems
Graphical Interpretation (3 of 3)

Figure 10.7
10-11
Constrained Optimization in Nonlinear Problems
Characteristics

■ Unlike linear programming, solution is often not on the

boundary of the feasible solution space.
■ Cannot simply look at points on the solution space boundary but
must consider other points on the surface of the objective
function.
■ This greatly complicates solution approaches.
■ Solution techniques can be very complex.

10-12
Western Clothing Problem
Solution Using Excel (1 of 3)

Exhibit 10.1
10-13
Western Clothing Problem
Solution Using Excel (2 of 3)

Exhibit 10.2

10-14
Western Clothing Problem
Solution Using Excel (3 of 3)

Exhibit 10.3
10-15
Beaver Creek Pottery Company Problem
Solution Using Excel (1 of 6)

Maximize Z = $(4 - 0.1x1)x1 + (5 - 0.2x2)x2

subject to:
x1 + 2x2 = 40
Where:
x1 = number of bowls produced
x2 = number of mugs produced

4 – 0.1x1 = profit ($) per bowl

5 – 0.2x2 = profit ($) per mug

10-16
Beaver Creek Pottery Company Problem
Solution Using Excel (2 of 6)

Exhibit 10.4
Copyright © 2010 Pearson Education, Inc. Publishing as Prentice Hall 10-17
Beaver Creek Pottery Company Problem
Solution Using Excel (3 of 6)

Exhibit 10.5
10-18
Beaver Creek Pottery Company Problem
Solution Using Excel (4 of 6)

Exhibit 10.6
10-19
Beaver Creek Pottery Company Problem
Solution Using Excel (5 of 6)

Exhibit 10.7

10-20
Beaver Creek Pottery Company Problem
Solution Using Excel (6 of 6)

Exhibit 10.8
10-21
Lagrange Multiplier
■ Lagrange multiplier, provides the dual value of the labor resource
■ Reflects the rate of change of the objective function with respect to
a small change in the constraint's right-hand side.
■ If you relax or tighten the constraint slightly:
■ How much the optimal value of the objective function would
increase (for maximization problems)
■ How much the optimal value of the objective function would
decrease (for minimization problems)

10-22
Interpretation of Values in the Sensitivity Report
■ Positive Lagrange multiplier: Indicates that
■ Increasing the constraint boundary would improve the objective
function value
■ Constraint is binding, meaning it directly impacts the solution.
■ provides the dual value of the labor resource
■ Zero Lagrange multiplier: Indicates that
■ Constraint is non-binding at the optimum (solution doesn’t
change if the constraint boundary is altered slightly)
■ The constraint is inactive at the optimal point
■ Objective function is unaffected by changes to this constraint.

10-23
Interpretation of Values in the Sensitivity Report
■ Negative Lagrange multiplier: Indicates that
■ Relaxing the constraint (e.g., loosening a minimum requirement)
would degrade the objective function)
■ Conversely, tightening the constraint could improve the objective.
■ In our example:
■ If the quantity of labor hours is increased from 40 to 41 hours, the
value of Z will increase by $0.33—from $70.42 to $70.75.

10-24
A Nonlinear Programming Model with Multiple
Constraints
Western Clothing Company example presented earlier, except now the
company produces two kinds of jeans, designer and straight-leg, and
production is subject to resource constraints for denim cloth, cutting
time, and sewing time. The company sells its jeans to several upscale
clothing store chains, and sales demand is dependent on the price at
which the company sells the jeans. The demand for designer jeans
(𝑥1 ) and the demand for straight-leg jeans (𝑥2 ) are defined by the
following relationships:

10-25
Western Clothing Company Problem
Solution Using Excel (1 of 4)
Maximize Z = (p1 - 12)x1 + (p2 - 9)x2
subject to:
2x1 + 2.7x2  6,000
3.6x1 + 2.9x2  8,500
7.2x1 + 8.5x2  15,000
where:
x1 = 1,500 - 24.6p1
x2 = 2,700 - 63.8p2
p1 = price of designer jeans
p2 = price of straight jeans
• Decision variables for this problem are 𝑝1 𝑎𝑛𝑑 𝑝2 , 𝑛𝑜𝑡 𝑥1 𝑎𝑛𝑑 𝑥2
• 𝑥1 & 𝑥2 , are functions of price and, thus, are dependent variables
• We did not substitute the functional relationships for 𝑥1 & 𝑥2 into
the objective function
10-26
Western Clothing Company Problem
Solution Using Excel (2 of 4)

Exhibit 10.9
10-27
Western Clothing Company Problem
Solution Using Excel (3 of 4)

Exhibit 10.10
10-28
Western Clothing Company Problem
Solution Using Excel (4 of 4)

Exhibit 10.11
10-29
Facility Location
■ Positive Lagrange multiplier: Indicates that
■ Increasing the constraint boundary would improve the objective
function value
■ Constraint is binding, meaning it directly impacts the solution.
■ provides the dual value of the labor resource
■ Zero Lagrange multiplier: Indicates that
■ Constraint is non-binding at the optimum (solution doesn’t
change if the constraint boundary is altered slightly)
■ The constraint is inactive at the optimal point
■ Objective function is unaffected by changes to this constraint.

10-30
Facility Location Example Problem
Problem Definition and Data (1 of 2)
• Centrally locate a facility that serves several customers or other
facilities in order to minimize distance or miles traveled (d) between
facility and customers.
• Distance formula for the straight line distance between two points
on a set of x, y coordinates is used (hypotenuse of a right triangle)

𝑑 = 𝑥𝑖 − 𝑥 2 + 𝑦𝑖 − 𝑦 2

Where:
(𝑥, 𝑦) = coordinates of proposed facility
(𝑥𝑖 , 𝑦𝑖 ) = coordinates of customer or location facility 𝑖

10-31
Facility Location Example Problem
Problem Definition and Data (1 of 2)
• Consider, for example, the Clayton County Rescue Squad and
Ambulance Service, which serves five rural towns, Abbeville,
Benton, Clayton, Dunning, and Eden. The rescue squad wants
to construct a centralized facility and garage to minimize its
total annual travel mileage to the towns.
• The locations of the five towns in terms of their graphical 𝑥, 𝑦
coordinates, measured in miles relative to the point 𝑥 = 0, 𝑦 = 0,
and the expected number of annual trips the squad will have to
make to each town are:

10-32
Facility Location Example Problem
Problem Definition and Data (1 of 2)
• The objective of the problem is to determine a set of coordinates
(𝑥, 𝑦) for the rescue squad facility that minimizes the total miles
traveled to the town
 𝑑𝑖𝑡𝑖
Where:
𝑑𝑖 = 𝑑𝑖𝑠𝑡𝑎𝑛𝑐𝑒 𝑡𝑜 𝑡𝑜𝑤𝑛 𝑖
𝑡𝑖 = 𝑎𝑛𝑛𝑢𝑎𝑙 𝑡𝑟𝑖𝑝𝑠 𝑡𝑜 𝑡𝑜𝑤𝑛 𝑖

10-33
Facility Location Example Problem
Solution Using Excel

Exhibit 10.12
10-34
Facility Location Example Problem
Solution Map

Figure 10.8 Rescue Squad Facility Location

10-35
Investment Portfolio Selection Example Problem
Definition and Model Formulation

■ A classic example: Investment portfolio selection

■ Based on assumption that most investors are concerned with 2 factors
1. Return on investment
2. Risk
Objective of the portfolio selection model is to:
■ Minimize some measure of portfolio Risk
■ While achieving some specified minimum return on the total portfolio
investment.
■ Risk: is reflected by the variability in the value of the investment
■ Here variance in the return on investment is the measure of Risk

10-36
Investment Portfolio Selection Example Problem
Definition and Model Formulation (1 of 2)

■ Covariance, (a measure of correlation), is also used to reflect risk.

■ Individual investment returns within a portfolio typically exhibit
statistical dependence.
■ Over time, the returns of any two stocks may exhibit positive or
negative correlation
■ That means two stocks of the same general type will go up or down
together.
■ To adjust for this possible correlation, investors often attempt to
diversify their portfolios.
■ To reflect the risk associated with not diversifying, the model includes
covariance.

10-37
Investment Portfolio Selection Example Problem
Definition and Model Formulation (2 of 2)
■ The minimization of portfolio risk, as measured by the portfolio
variance, is the model objective
■ The variance, S, on the annual return from the portfolio is
determined by the following formula:
𝑺 = 𝒙𝟏𝟐𝒔𝟏𝟐 + 𝒙𝟐𝟐𝒔𝟐𝟐 + … + 𝒙𝒏𝟐𝒔𝒏𝟐 + ෍ 𝒙𝒊𝒙𝒋𝒓𝒊𝒋𝒔𝒊𝒔𝒋
i≠j
𝒙𝒊, 𝒙𝒋 = the proportion of money invested in investments 𝒊 𝒐𝒓 𝒋
𝒔𝒊𝟐 = the variance for investment 𝒊
𝒓𝒊𝒋 = the correlation between returns on investments 𝒊 𝒂𝒏𝒅 𝒋
𝒔𝒊, 𝒔𝒋 = the std. dev. of returns for investments 𝒊 𝒂𝒏𝒅 𝒋

▪ First part is a measure of variance, and the second part is a

measure of the covariance
10-38
Investment Portfolio Selection Example Problem
Definition and Model Formulation (2 of 2)
▪ In many cases, parameters in this equation may be estimates or
sample values (the sample variance, sample covariance, etc.).
▪ The general formulation of the portfolio selection model is as
follows.
▪ The investor desires to achieve a minimum expected annual return
from the portfolio, which is formulated as a model constraint as:
𝒓𝟏𝒙𝟏 + 𝒓𝟐𝒙𝟐 + … + 𝒓𝒏𝒙𝒏  𝒓𝒎
where:
𝑟𝑖 = expected annual return on investment 𝑖
𝑥𝑖 = proportion (fraction) of money invested in investment 𝑖
𝑟𝑚 = the minimum desired annual return from the portfolio

▪ A second constraint specifies that all the money is invested:

𝒙𝟏 + 𝒙𝟐 + … 𝒙𝒏 = 𝟏. 𝟎
10-39
Investment Portfolio Selection Example
▪ Jessica Todd has identified four stocks she wants to include in her
investment portfolio. She wants a total annual return of at least
.11. From historical data, she has estimated the average returns and
variances for the four investments as follows:
Stock (xi) Annual Return (ri) Variance (si)
Altacam .08 .009
Bestco .09 .015
Com.com .16 .040
Delphi .12 .023

▪ She has also estimated the covariances between the stocks, as follows
(on next slide):

10-40
Investment Portfolio Selection Example Problem
Solution Using Excel (1 of 5)

Stock (xi) Annual Return (ri) Variance (si)

Altacam .08 .009
Bestco .09 .015
Com.com .16 .040
Delphi .12 .023

10-41
Investment Portfolio Selection Example Problem
Solution Using Excel (2 of 5)
Four stocks, desired annual return of at least 0.11.
Minimize
Z = S = x12(.009) + x22(.015) + x32(.040) + X42(.023)
+ x1x2 (.4)(.009)1/2(0.015)1/2 + x1x3(.3)(.009)1/2(.040)1/2 +
x1x4(.6)(.009)1/2(.023)1/2 + x2x3(.2)(.015)1/2(.040)1/2 +
x2x4(.7)(.015)1/2(.023)1/2 + x3x4(.4)(.040)1/2(.023)1/2 +
x2x1(.4)(.015)1/2(.009)1/2 + x3x1(.3)(.040)1/2 + (.009)1/2 +
x4x1(.6)(.023)1/2(.009)1/2 + x3x2(.2)(.040)1/2(.015)1/2 +
x4x2(.7)(.023)1/2(.015)1/2 + x4x3(.4)(.023)1/2(.040)1/2
subject to:
.08x1 + .09x2 + .16x3 + .12x4  0.11
x1 + x2 + x3 + x4 = 1.00
xi  0
10-42
Investment Portfolio Selection Example Problem
Solution Using Excel (3 of 5)

Exhibit 10.13
10-43
Investment Portfolio Selection Example Problem
Solution Using Excel (4 of 5)

Exhibit 10.14
10-44
Investment Portfolio Selection Example Problem
Solution Using Excel (5 of 5)

Exhibit 10.15
10-45
Hickory Cabinet and Furniture Company
Example Problem and Solution (1 of 2)

Model:
Maximize Z = $280x1 - 6x12 + 160x2 - 3x22
subject to:
20x1 + 10x2 = 800 board ft.
Where:
x1 = number of chairs
x2 = number of tables

10-46
Hickory Cabinet and Furniture Company
Example Problem and Solution (2 of 2)

10-47
 Problem 1

Using Excel Solver, determine the price to maximize profit

𝑴𝒂𝒙 𝒁 = 𝒗𝒑 − 𝑪𝒇 − 𝒗𝑪𝒗

10-48
 Problem 5

𝑴𝒂𝒙𝒊𝒎𝒊𝒛𝒆 𝒁 = $(𝟓𝟎 − 𝟎. 𝟑𝒙𝟏)𝒙𝟏 + (𝟒𝟐. − 𝟎. 𝟔𝒙𝟐)𝒙𝟐

subject to:
𝒙𝟏 + 𝟏. 𝟐𝒙𝟐 ≤ 𝟏𝟐𝟎
Where:
x1 = Demand of Tire A
x2 = Demand of Tire B

𝟓𝟎 − 𝟎. 𝟑𝒙𝟏 = profit ($) per Tire A

𝟒𝟐. − 𝟎. 𝟔𝒙𝟐 = profit ($) per Tire B

10-49
Problem 14

10-50
10-51
10-52
Problem 18

10-53
10-54
Problem 21

10-55
10-56