Integer Linear Programming

Rajendra Desai, XIME, 2012

Slide 1

Integer Linear Programming

 

 

Types of Integer Linear Programming Models Graphical and Computer Solutions for an AllAllInteger Linear Program Applications Involving 0-1 Variables 0Modeling Flexibility Provided by 0-1 Variables 0-

Rajendra Desai, XIME, 2012

Slide 2

Types of Integer Programming Models



An LP in which all the variables are restricted to be integers is called an all-integer linear program (ILP). allThe LP that results from dropping the integer requirements is called the LP Relaxation of the ILP. If only a subset of the variables are restricted to be integers, the problem is called a mixed-integer linear mixedprogram (MILP). Binary variables are variables whose values are restricted to be 0 or 1. If all variables are restricted to be 0 or 1, the problem is called a 0-1 or binary integer linear program. program.

Rajendra Desai, XIME, 2012

Slide 3

Example: All-Integer LP All

Consider the following all-integer linear program: allMax s.t. 3x 2x 3x1 + 2x2 3x 3x1 + x2 < 9 x1 + 3x2 < 7 3x -x 1 + x 2 < 1 x1, x2 > 0 and integer

Rajendra Desai, XIME, 2012

Slide 4

Example: All-Integer LP All

LP Relaxation Solving the problem as a linear program ignoring the integer constraints, the optimal solution to the linear program gives fractional values for both x1 and x2. From the graph on the next slide, we see that the optimal solution to the linear program is: x1 = 2.5, x2 = 1.5, Max 3x1 + 2x2 = 10.5 3x 2x

Rajendra Desai, XIME, 2012

Slide 5

Example: All-Integer LP All

LP Relaxation x2
5 4 3 2 1  x1 + x2 < 1 3x1 + x2 < 9 Max 3x1 + 2x2 3x 2x LP Optimal (2.5, 1.5) x1 + 3x2 < 7 3x

x1
Slide 6

Rajendra Desai, XIME, 2012

Example: All-Integer LP All

Rounding Up If we round up the fractional solution (x1 = 2.5, (x x2 = 1.5) to the LP relaxation problem, we get x1 = 3 and x2 = 2. From the graph on the next slide, we see that this point lies outside the feasible region, making this solution infeasible.

Rajendra Desai, XIME, 2012

Slide 7

Example: All-Integer LP All

Rounded Up Solution x2
5 4 3 2 1  x1 + x2 < 1 3x1 + x2 < 9 Max 3x1 + 2x2 3x 2x ILP Infeasible (3, 2) LP Optimal (2.5, 1.5) x1 + 3x2 < 7 3x

x1
Slide 8

Rajendra Desai, XIME, 2012

Example: All-Integer LP All

Rounding Down By rounding the optimal solution down to x1 = 2, x2 = 1, we see that this solution indeed is an integer solution within the feasible region, and substituting in the objective function, it gives 3x1 + 2x2 = 8. 3x 2x We have found a feasible all-integer solution, but allhave we found the OPTIMAL all-integer solution? all--------------------The answer is NO! The optimal solution is x1 = 3 and x2 = 0 giving 3x1 + 2x2 = 9, as evidenced in the 3x 2x next two slides.

Rajendra Desai, XIME, 2012

Slide 9

Example: All-Integer LP All

Complete Enumeration of Feasible ILP Solutions There are eight feasible integer solutions to this problem: x1 x2 3x1 + 2x2 2x 1. 2. 3. 4. 5. 6. 7. 8. 0 1 2 3 0 1 2 1 0 0 0 0 1 1 1 2 0 3 6 9 2 5 8 7

optimal solution

Rajendra Desai, XIME, 2012

Slide 10

Example: All-Integer LP All

Optimal All-Integer Solution Allx2

5 4 3 2 1  x1 + x2 < 1 3x1 + x2 < 9 Max 3x1 + 2x2 3x 2x

ILP Optimal (3, 0)

x1 + 3x2 < 7 3x

x1
Slide 11

Rajendra Desai, XIME, 2012

Example: Shyams Tailoring

Shyams Tailoring has five idle tailors and four custom garments to make. The estimated time (in hours) it would take each tailor to make each garment is shown in the next slide. (An 'X' in the table indicates an unacceptable tailor-garment assignment.) tailorTailor 2 3 4 5 23 20 21 18 14 X 12 10 8 11 X 9 20 20 18 21

Garment Wedding gown Clown costume Admiral's uniform Bullfighter's outfit

1 19 11 12 X

Rajendra Desai, XIME, 2012

Slide 12

Example: Shyams Tailoring

Formulate an integer program for determining the tailor-garment assignments that minimize tailorthe total estimated time spent making the four garments. No tailor is to be assigned more than one garment and each garment is to be worked on by only one tailor. -------------------This problem can be formulated as a 0-1 integer 0program. The LP solution to this problem will automatically be integer (0-1). (0-

Rajendra Desai, XIME, 2012

Slide 13

Example: Shyams Tailoring



Define the decision variables xij = 1 if garment i is assigned to tailor j = 0 otherwise. Number of decision variables = [(number of garments)(number of tailors)] - (number of unacceptable assignments) = [4(5)] - 3 = 17

Rajendra Desai, XIME, 2012

Slide 14

Example: Shyams Tailoring



Define the objective function Minimize total time spent making garments: Min 19x11 + 23x12 + 20x13 + 21x14 + 18x15 + 11x21 19x 23x 20x 21x 18x 11x + 14x22 + 12x24 + 10x25 + 12x31 + 8x32 + 11x33 14x 12x 10x 12x 8x 11x + 9x35 + 20x42 + 20x43 + 18x44 + 21x45 20x 20x 18x 21x

Rajendra Desai, XIME, 2012

Slide 15

Example: Shyams Tailoring



Define the Constraints Exactly one tailor per garment: 1) x11 + x12 + x13 + x14 + x15 = 1 2) x21 + x22 + x24 + x25 = 1 3) x31 + x32 + x33 + x35 = 1 4) x42 + x43 + x44 + x45 = 1

Rajendra Desai, XIME, 2012

Slide 16

Example: Shyams Tailoring



Define the Constraints (continued) No more than one garment per tailor: 5) x11 + x21 + x31 < 1 6) x12 + x22 + x32 + x42 < 1 7) x13 + x33 + x43 < 1 8) x14 + x24 + x44 < 1 9) x15 + x25 + x35 + x45 < 1 Nonnegativity: xij > 0 for i = 1, . . ,4 and j = 1, . . ,5

Rajendra Desai, XIME, 2012

Slide 17

Modeling Flexibility Provided by 0-1 Variables 0

When xi and xj represent binary variables designating whether projects i and j have been completed, the following special constraints may be formulated:

At most k out of n projects will be completed:

7xj < k j Project j is conditional on project i: xj - xi < 0 Project i is a corequisite for project j: xj - xi = 0 Projects i and j are mutually exclusive: exclusive: xi + xj < 1

Rajendra Desai, XIME, 2012

Slide 18

Example: Metropolitan Microwaves

Metropolitan Microwaves is planning to expand its sales operation by offering other electronic appliances. The company has identified seven new product lines it can carry. Relevant information about each line follows on the next slide.

Rajendra Desai, XIME, 2012

Slide 19

Example: Metropolitan Microwaves

Initial Floor Space Exp. Rate Invest. (Sq.Ft.) (Sq.Ft.) of Return 6,000 12,000 20,000 14,000 15,000 2,000 32,000 125 150 200 40 40 20 100 8.1% 9.0 11.0 10.2 10.5 14.1 13.2

Product Line 1. 2. 3. 4. 5. 6. 7. TV/VCRs TVs Projection TVs VCRs DVD Players Video Games Home Computers

Rajendra Desai, XIME, 2012

Slide 20

Example: Metropolitan Microwaves

Metropolitan has decided that they should not stock projection TVs unless they stock either TV/VCRs or TVs. Also, they will not stock both VCRs and DVD players, and they will stock video games if they stock TVs. Finally, the company wishes to introduce at least three new product lines. If the company has 45,000 to invest and 420 sq. ft. of floor space available, formulate an integer linear program for Metropolitan to maximize its overall expected return.

Rajendra Desai, XIME, 2012

Slide 21

Example: Metropolitan Microwaves



Define the Decision Variables xj = 1 if product line j is introduced; = 0 otherwise. where: Product line 1 = TV/VCRs Product line 2 = TVs Product line 3 = Projection TVs Product line 4 = VCRs Product line 5 = DVD Players Product line 6 = Video Games Product line 7 = Home Computers

Rajendra Desai, XIME, 2012

Slide 22

Example: Metropolitan Microwaves



Define the Decision Variables xj = 1 if product line j is introduced; = 0 otherwise. Define the Objective Function Maximize total expected return: Max .081(6000)x1 + .09(12000)x2 + .11(20000)x3 .081(6000)x .09(12000)x .11(20000)x + .102(14000)x4 + .105(15000)x5 + .141(2000)x6 .102(14000)x .105(15000)x .141(2000)x + .132(32000)x7 .132(32000)x

Rajendra Desai, XIME, 2012

Slide 23

Example: Metropolitan Microwaves



Define the Constraints 1) Money: 6x1 + 12x2 + 20x3 + 14x4 + 15x5 + 2x6 + 32x7 < 45 12x 20x 14x 15x 2x 32x 2) Space: 125x +150x +200x +40x +40x +20x +100x 125x1 +150x2 +200x3 +40x4 +40x5 +20x6 +100x7 < 420 3) Stock projection TVs only if stock TV/VCRs or TVs: x1 + x2 > x3 or x1 + x2 - x3 > 0

Rajendra Desai, XIME, 2012

Slide 24

Example: Metropolitan Microwaves



Define the Constraints (continued) 4) Do not stock both VCRs and DVD players: x4 + x5 < 1 5) Stock video games if they stock TV's: x2 - x6 > 0 6) Introduce at least 3 new lines: x1 + x2 + x3 + x4 + x5 + x6 + x7 > 3 7) Variables are 0 or 1: xj = 0 or 1 for j = 1, , , 7

Rajendra Desai, XIME, 2012

Slide 25

Example: Metropolitan Microwaves



Partial Spreadsheet Showing Problem Data

A B C X2 12 150 1 0 1 1 1080 1 2 Constraints X1 3 #1 6 4 #2 125 5 #3 1 6 #4 0 7 #5 0 8 #6 1 9 Obj.Func.Coeff. 486 D E F LHS Coefficients X3 X4 X5 20 14 15 200 40 40 -1 0 0 0 1 1 0 0 0 1 1 1 2200 1428 1575 G X6 2 20 0 0 -1 1 282 H I

X7 RHS 32 45 100 420 0 0 0 1 0 0 1 3 4224

Rajendra Desai, XIME, 2012

Slide 26

Example: Metropolitan Microwaves



Partial Spreadsheet Showing Example Formulas

12 13 14 15 16 17 18 19 20 21 22 B C D X1 X2 X3 Dec.Values 0 0 0 Maximized Total Expected Return Constraints Money Space TVs VCRs Video Lines LHS 0 0 0 0 0 0 A E X4 0 F X5 0 0 G X6 0 H X7 0

<= <= >= <= >= >=

RHS 45 420 0 1 0 3

Rajendra Desai, XIME, 2012

Slide 27

Example: Metropolitan Microwaves



Optimal Solution
12 13 14 15 16 17 18 19 20 21 22 B C D X1 X2 X3 Dec.Values 1 0 1 Maximized Total Expected Return Constraints Money Space TVs VCRs Video Lines LHS 41 365 0 1 0 3 A E X4 0 F X5 1 4261 G X6 0 H X7 0

<= <= >= <= >= >=

RHS 45 420 0 1 0 3

Rajendra Desai, XIME, 2012

Slide 28

Example: Metropolitan Microwaves



Optimal Solution Introduce: TV/VCRs, Projection TVs, and DVD Players Do Not Introduce: TVs, VCRs, Video Games, and Home Computers Total Expected Return: Return: $4,261

Rajendra Desai, XIME, 2012

Slide 29

Cautionary Note About Sensitivity Analysis



Sensitivity analysis often is more crucial for ILP problems than for LP problems. A small change in a constraint coefficient can cause a relatively large change in the optimal solution. Recommendation: Resolve the ILP problem several times with slight variations in the coefficients before choosing the best solution for implementation.

Rajendra Desai, XIME, 2012

Slide 30