Beyond Linear Programming: Mathematical Programming Extensions CHAPTER 7 and 17 (on CD)

Integer Programming Goal Programming Non-Linear Programming

ADM2302 -- Rim Jaber

Learning Objectives

Formulate integer programming (IP) models. Set up and solve IP models using Excels Solver. Understand difference between general integer and binary integer variables

Understand use of binary integer variables in formulating problems involving fixed (or setup) costs.
Formulate goal programming problems and solve them using Excels Solver.

ADM2302 -- Rim Jaber

Integer Programming

ADM2302 -- Rim Jaber

Integer Programming (IP)

Constraints and Objective function similar to the LP model

Difference:

one or more of the decisions variables has to take on an integer value in the final solution.

Example of IP: Harison Electric Company

ADM2302 -- Rim Jaber 4

Example 1: Harison Electric Company

The Harison Electric Company, located in Chicagos old town area, Produces two products popular with home renovators: old-fashion Chandelier and ceiling fans. Both the chandeliers and fans require a two-step production process involving wiring and assembly. It takes about 2 hours to wire each chandelier, and 3 hours to wire a ceiling fan. Final assembly of the chandeliers and fans requires 6 and 5 hours, respectively. The production capability is such that 12 hours of wiring time and 30 hours of assembly time are available. If each chandelier produced nets the firm $600 and each fan $700. Harisons production mix decision can be formulated using LP as follows:
ADM2302 -- Rim Jaber 5

Harrison Electric Problem

Maximize Profit = $ 600X1 + 700X2 Subject to 2X1 + 3X2 <= 12 (wiring hours) 6X1 + 5X2 <= 30 (assembly hours) X1, X2 >= 0 X1 = number of chandeliers produced X2 = number of ceiling fans produced
ADM2302 -- Rim Jaber 6

X2 6 5 4 3 2 1 0

Harrison Electric Problem

6X1 + 5X2 30
+ = possible Integer Solution

+
+ + +

Optimal LP Solution (X1 = 3 3/4, X2 = 1 1/2 Profit = $3,300)

+
1

+
2

+
3

+
4 5 6

2X1 + 3X2 12 X1

ADM2302 -- Rim Jaber

Comments on Reaching Integer Solution

Rounding off is one way to reach integer solution values, but it often does not yield the best solution.

Problems of Rounding:

ADM2302 -- Rim Jaber

Integer Solution to the Harrison Electric Company Problem

Chandeliers (X1)
0 1 2 3 4 5 0 1 2 3 4 0 1 2 3 0 1 0

Ceiling Fans(X2)
0 0 0 0 0 0 1 1 1 1 1 2 2 2 2 3 3 4

Profit: $600X1 + $700X2

$0 600 1,200 1,800 2,400 3,000 700 1,300 1,900 2,500 3,100 1,400 2,000 2,600 3,200 2.100 2,700 2,800

Solution if rounding off is used Optimal Solution to integer Programming Problem 9

Comments on Reaching Integer Solution-Contd

An IP solution can never be better than the solution to the same LP problem. The integer problem is usually worse in terms of higher cost or lower profit. Although enumeration is feasible for some small integer programming problems, it can be difficult or impossible for large ones.
ADM2302 -- Rim Jaber 10

Integer Programming Techniques

Cutting Plane Method

A means of adding one or more constraints to LP problems to help produce an optimum integer solution
An algorithm for solving all-integer and mixedinteger LP and assignment problems. It divides the set of feasible solutions into subsets that are examined systematically. Solver uses this method
ADM2302 -- Rim Jaber 11

Branch and Bound Method

Formulating and Solving Spreadsheet Models for Integer Programming Problems

Similar to LP except we need to include the constraints that the decisions variables need to be integer. Changing Cells = integer Solver Options

Maximum Time Allowed. Max Time option set to 100 seconds default value. Tolerance of the Optimal Solution. Tolerance option set at 5% default value.
ADM2302 -- Rim Jaber 12

Refer to BIP.xsl

Types of Integer Programming Problems

Pure Integer Programming

Mixed Integer Programming

all variables must have integer solutions some, but not all variables have integer solutions all variables have values of 0 or 1

Binary (0-1) Integer Programming

Mixed binary integer programming problems.

Some decision variables are binary, and other decision variables are either general integer or continuous valued.
ADM2302 -- Rim Jaber 13

Binary Integer Programming

ADM2302 -- Rim Jaber

14

Binary Integer Programming

Binary Integer Programming (BIP): the model for a BIP is identical to that for a LP problem except that the nonnegativity constraints for at least some of the variables are binary variables.

Pure PIB problem Mixed PIB problem

ADM2302 -- Rim Jaber 15

Binary Integer Programming (Contd)

Binary Variables:

A variable whose only possible values are 0 or 1.

Used when dealing with yes-or-no decisions
Example 2

OR used to help formulate the model constraints and/or objective function (do not represent a yes-or-no decision)

Binary (zero-one) variables are defined as Y= 1 if an event takes place 0 if an event doesnt take place
ADM2302 -- Rim Jaber 16

Called auxiliary binary variable Example 3

BIP: Example 2 -- Employee Scheduling Application

The Department head of a management science department at a major Midwestern university will be scheduling faculty to teach courses during the coming autumn term. Four core courses need to be covered. The four courses are at the UG, MBA, MS, and Ph.D. levels. Four professors will be assigned to the courses, with each professor receiving one of the courses. Student evaluations of professors are available from previous terms. Based on a rating scale of 4 (excellent), 3 (very good), 2 (average), 1(fair), and 0(poor), the average student evaluations for each professor are shown: ADM2302 -- Rim Jaber 17

Professor D does not have a Ph.D. and cannot be assigned to teach the Ph.D.-level course. If the department head makes teaching assignments based on maximizing the student evaluation ratings over all four courses, what staffing assignments should be made?
Course

Professor
A B C D

UG 2.8 3.2 3.3 3.2

MBA 2.2 3.0 3.2 2.8

MS
3.3 3.6 3.5 2.5

Ph.D.
3.0 3.6 3.5 18

ADM2302 -- Rim Jaber

BIP Example 2: Solution

Excel Solution: BIP.xsl

ADM2302 -- Rim Jaber

19

Example 3
The Research and Development Division of the Progressive Company has been developing four possible new product lines. Management must now make a decision as to which of these Four products actually will be produced and at what levels.
Therefore, a management science study has been requested to Find the most profitable product mix.

A substantial cost is associated with beginning the production Of any product, as given in the first row of the following table.
Managements objective is to find the product mix that maximizes 20 The total profit (total net revenue minus start up costs).

Product
1 Start-up-cost Marginal revenue Maximum Production 2 3 4

$50,000 $40,000
$70 1000 $60 3000

$70,000
$90 2000

$60,000
$80 1000

Let the Integer decisions variables x1, x2, x3, and x4 be the total number of units produced of products 1, 2, 3 and 4, respectively. Management has imposed the following policy constraints on these variables:
ADM2302 -- Rim Jaber 21

No more than two of the products can be produced If product 3 or 4 is produced then product 1 or 2 must be produced Either 5x1+3x2+6x3+4x4 <= 6,000 Or Either 4x1+6x2+3x3+5x4 <= 6,000 Use Auxiliary binary variables to formulate and solve the mixed BIP model.

ADM2302 -- Rim Jaber 22

No more than two of the products can be produced

Yi = 1 if product i is produced 0 if product i is not produced Y1 + Y2 + Y3 + Y4 <= 2 Extension of the Mutually Exclusive Alternatives Constraints

Mutually exclusive constraint is a constraint requiring that the sum of two or more binary variables be less than or equal to 1. Thus if one of the variables is equal to 1 , the others must equal zero.
ADM2302 -- Rim Jaber 23

If product 3 or 4 is produced then product 1 or 2 must be produced Contingent Decision (Conditional Decision) : It can be yes only if a certain other yes-or-no decision is yes. Y3 <= Y1 + Y2 Y 4 <= Y1 + Y2 Contingency Constraints: it involves binary variables that do not allow certain variables to equal 1 unless certain other variables are equal to 1

ADM2302 -- Rim Jaber 24

Either 5x1+3x2+6x3+4x4 <= 6,000 Or Either 4x1+6x2+3x3+5x4 <= 6,000

Either-or constraints are not allowed in Linear or Integer Programming

A pair of constraints such that either one can be chosen to be observed and then the other one would be ignored

Use binary variables to reformulate this model into a standard format, in order to be able to find the optimal solution. Y= 0 if 5x1+3x2+6x3+4x4 <= 6,000 must hold Or 1 if 4x1+6x2+3x3+5x4 <= 6,000 must hold 5x1+3x2+6x3+4x4 <= 6,000 + 1,000,000Y 4x1+6x2+3x3+5x4 <= 6,000 + 1,000,000(1-Y)
ADM2302 -- Rim Jaber 25

Production Cost: Setup Costs (=Start-up Costs) + Variable Costs

Objective function Max Z = $70x1 + 60x2 + 90x3 +80x4

We need to subtract from this expression each setup cost if the corresponding product will be produced, but we should not subtract the setup cost if the product will not be produced.

ADM2302 -- Rim Jaber

26

Setup Costs (contd)

Y1 = Y2 = Y3 = Y4 =

1 0 1 0 1 0 1 0

if if if if if if if if

x1 x1 x2 x2 x3 x3 x4 x4

is is is is is is is is

produced not produced produced not produced produced not produced produced not produced

27

Objective Function for the Model

The total Start-up Cost is: $50,000Y1 + $40,000Y2 + $70,000Y3 + $60,000Y4 Max Z = $70x1 + $60x2 + $90x3 +$80x4 $50,000Y1 - $40,000Y2 - $70,000Y3 $60,000Y4 Constraints: x1 <= 1000Y1 x2 <= 3000 Y2 x3 <= 2000 Y3 x4 <= 1000 Y4

ADM2302 -- Rim Jaber 28

Max Z = $70x1 + $60x2 + $90x3 +$80x4 $50,000Y1 - $40,000Y2 - $70,000Y3 - $60,000Y4 Subject to Y1 + Y2 + Y3 + Y4 <= 2 Y3 <= Y1 + Y2 Y 4 <= Y1 + Y2 5x1+3x2+6x3+4x4 <= 6,000 + 1,000,000Y 4x1+6x2+3x3+5x4 <= 6,000 + 1,000,00(1-Y) x1 <= 1000Y1 x2 <= 3000Y2 x3 <= 2000Y3 x4 <= 1000Y4 x1, x2, x3, x4 >= 0 and integer; Y,Y1,Y2,Y3,Y4 = 0, 1

The Mixed BIP Model Summary

Summary
Integer LP models have variety of applications including: capital budgeting problems (individual projects are represented as binary variables) facilities location problems (selecting a location is represented as a binary variable) airline crew scheduling problems (assigning crew to a particular flight is represented as a binary variable) knapsack problems (loading an item into a container is represented as a binary variable)
ADM2302 -- Rim Jaber 30