31
1 Lecture 5 Lecture 5 Set Packing Problems Set Packing Problems Set Partitioning Set Partitioning Problems Problems

1 Lecture 5 Set Packing Problems Set Partitioning Problems

Embed Size (px)

Citation preview

Page 1: 1 Lecture 5 Set Packing Problems Set Partitioning Problems

1

Lecture 5 Lecture 5 Set Packing ProblemsSet Packing Problems

Set Partitioning ProblemsSet Partitioning Problems

Page 2: 1 Lecture 5 Set Packing Problems Set Partitioning Problems

2

OutlineOutline

set packing problems set packing problems

set partitioning problemsset partitioning problems

Page 3: 1 Lecture 5 Set Packing Problems Set Partitioning Problems

3

Set Packing ProblemsSet Packing Problems

Page 4: 1 Lecture 5 Set Packing Problems Set Partitioning Problems

4

ContextContext

a set a set S S = {1, 2…, = {1, 2…, mm} }

a collection of subsets of a collection of subsets of SS,, , such that , such that

each subset carry a value each subset carry a value problem: to maximize the total value of problem: to maximize the total value of

subsets selected such that no element is subsets selected such that no element is

selected more than onceselected more than once

Page 5: 1 Lecture 5 Set Packing Problems Set Partitioning Problems

5

Set Packing ProblemsSet Packing Problems

S S = {1, 2, 3, 4, 5, 6} = {1, 2, 3, 4, 5, 6}

= {{1, 2, 5}, {1, 3}, {2, 4}, {3, 6}, {2, 3, 6}}= {{1, 2, 5}, {1, 3}, {2, 4}, {3, 6}, {2, 3, 6}} examples examples

{1, 2, 5}, {3, 6}: a pack{1, 2, 5}, {3, 6}: a pack

{1, 3}, {2, 4}: a pack{1, 3}, {2, 4}: a pack

{1, 2, 5}, {2, 3, 6}: not a pack{1, 2, 5}, {2, 3, 6}: not a pack

{2, 3, 6}: a pack{2, 3, 6}: a pack

assumption: every subset of value = 1assumption: every subset of value = 1

Page 6: 1 Lecture 5 Set Packing Problems Set Partitioning Problems

6

Set Packing ProblemsSet Packing Problems

S S = {1, 2, 3, 4, 5, 6} = {1, 2, 3, 4, 5, 6}

= {{1, 2, 5}, {1, 3}, {2, 4}, {3, 6}, {2, 3, 6}}= {{1, 2, 5}, {1, 3}, {2, 4}, {3, 6}, {2, 3, 6}}

: element 1

: element 2

: element 3

: element 4

: element 5

: element 6

1 2 3 4 5

1 2

1 3 5

2 4 5

3

1

2 4 5

max ,

. . 1,

1,

1,

1,

1,

1,

s t

i {0, 1}

set

1: set

2: set

3: set

4: set

5:

Property 9.4: maximization

with all constraints

Property 9.5: all RHS

coefficients = 1

Property 9.6: all matrix

coefficients = 0 or 1

1, if the th member of is in the pack,

0, otherwise.ii

Page 7: 1 Lecture 5 Set Packing Problems Set Partitioning Problems

7

Primal-Dual PairPrimal-Dual Pair

“An interesting observation is that the LP problem associated with a set packing problem with objective coefficients of 1 is the dual of the LP problem associated with a set covering problem with objective coefficients of 1.” (pp 192 of [7])

Page 8: 1 Lecture 5 Set Packing Problems Set Partitioning Problems

8

Primal-Dual PairPrimal-Dual Pair

special properties between the primal-dual pairspecial properties between the primal-dual pair

obj function of min obj function of min obj function of max obj function of max

unbounded primal unbounded primal infeasible dual infeasible dual

optimal primal optimal primal optimal dual optimal dual

same objective function valuesame objective function value

easy to deduce the optimal of one from the other easy to deduce the optimal of one from the other

possible to have infeasible primal and infeasible dualpossible to have infeasible primal and infeasible dual

1 1 2 2

11 1 12 2 1 1

1 1 2 2

max ... ,

. .

... ,

... ,

0.

n n

n n

m m mn n m

i

c x c x c x

s t

a x a x a x b

a x a x a x b

x

M M M M M

Primal (Dual)

1 1 2 2

11 1 21 2 1 1

1 1 2 2

min ... ,

. .

... ,

... ,

0.

m m

m n

n n mn n n

j

b y b y b y

s t

a y a y a y c

a y a y a y c

y

M M M M M

Dual (Primal)

Page 9: 1 Lecture 5 Set Packing Problems Set Partitioning Problems

9

Comments on Comments on Set Packing ProblemsSet Packing Problems

1 2 3 4 5

1 2

1 3 5

2 4 5

3

1

2 4 5

max ,

. . 1,

1,

1,

1,

1,

1,

s t

i 0

set packing problem with objective coefficients = 1

1 2 3 4 5 6

1 2 5

1 3

2 4

3 6

2 3 6

min ,

. .

1,

1,

1,

1,

1,

s t

i 0

set covering problem with objective coefficients = 1

dual of each other

The two LPs are dual of each other, though the S and in one problem are different from those of the other

problem.

Page 10: 1 Lecture 5 Set Packing Problems Set Partitioning Problems

10

Comments on Comments on Set Packing ProblemsSet Packing Problems

similar generalization as in set covering similar generalization as in set covering

problemsproblems weighted set packing problems: RHS weighted set packing problems: RHS

positive integers > 1positive integers > 1

generalized set packing problems: matrix generalized set packing problems: matrix

coefficients = 0 or coefficients = 0 or 1 1

Page 11: 1 Lecture 5 Set Packing Problems Set Partitioning Problems

11

Matching Problem: Matching Problem: A Special Type of Set Packing ProblemA Special Type of Set Packing Problem

matching: select the maximum numbers matching: select the maximum numbers

of arcs such that there is no overlapping of arcs such that there is no overlapping

of nodes involved of nodes involved (1, 2), (3, 6), (4, 5)(1, 2), (3, 6), (4, 5)

2

3

15

4

6

Page 12: 1 Lecture 5 Set Packing Problems Set Partitioning Problems

12

ExerciseExercise

formulate the matching formulate the matching

problem of the RHS problem of the RHS

network as a set packing network as a set packing

problem problem

2

3

15

4

6

Page 13: 1 Lecture 5 Set Packing Problems Set Partitioning Problems

13

Set Partitioning ProblemsSet Partitioning Problems

Page 14: 1 Lecture 5 Set Packing Problems Set Partitioning Problems

14

Set Partitioning ProblemsSet Partitioning Problems

to cover all the members of to cover all the members of SS by elements of by elements of

without overlapping without overlapping

S S = {1, 2, 3, 4, 5} = {1, 2, 3, 4, 5}

= {{1, 2}, {1, 3, 5}, {2, 4, 5}, {3}, {1}, {4, 5}} = {{1, 2}, {1, 3, 5}, {2, 4, 5}, {3}, {1}, {4, 5}}

e.g., {1, 2}, {3}, and {4, 5} form a partitione.g., {1, 2}, {3}, and {4, 5} form a partition

Page 15: 1 Lecture 5 Set Packing Problems Set Partitioning Problems

15

Set Partitioning ProblemsSet Partitioning Problems

1, if the th member of is in the partition,

0, otherwise.ii

1 2 5

1 3

2 4

3 6

2 3 6

1,

1,

1,

1,

1.

i {0, 1}

either maximization or minimization is all right

S S = {1, 2, 3, 4, 5} = {1, 2, 3, 4, 5}

= {{1, 2}, {1, 3, 5}, {2, 4, 5}, {3}, {1}, {4, 5}} = {{1, 2}, {1, 3, 5}, {2, 4, 5}, {3}, {1}, {4, 5}}

Page 16: 1 Lecture 5 Set Packing Problems Set Partitioning Problems

16

Equivalence Between Set Packing Problem Equivalence Between Set Packing Problem and Set Partitioning Problemand Set Partitioning Problem

1 2 3 4 5

1 2

1 3 5

2 4 5

3

1

2 4 5

max ,

. . 1,

1,

1,

1,

1,

1,

s t

i {0, 1}

Set Packing Problem

1 2 3 4 5

1 2 6

1 3 5 7

2 4 5 8

3 9

1 10

2 4 5 11

max ,

. . 1,

1,

1,

1,

1,

1,

s t

i {0, 1}, 1 to 11

Set Partitioning Problem

Page 17: 1 Lecture 5 Set Packing Problems Set Partitioning Problems

17

Equivalence Between Set Packing Problem Equivalence Between Set Packing Problem and Set Partitioning Problemand Set Partitioning Problem

illustrate the transformation of the first illustrate the transformation of the first

equality constraint into an equality constraint into an -constraint-constraint

introduce a dummy variable introduce a dummy variable to the first to the first

constraintconstraint

i {0, 1}

Set Partitioning Problem

1 2 5

1 3

2 4

3 6

2 3 6

1,

1,

1,

1,

1.

max 1 + 2 + 3 + 4 + 5 + 6,

Page 18: 1 Lecture 5 Set Packing Problems Set Partitioning Problems

18

Equivalence Between Set Packing Problem Equivalence Between Set Packing Problem and Set Partitioning Problemand Set Partitioning Problem

i {0, 1}

Set Partitioning Problem

1 2 5

1 3

2 4

3 6

2 3 6

1,

1,

1,

1,

1.

max 1 + 2 + 3 + 4 + 5 + 6,

i, {0, 1}

1 2 5

1 3

2 4

3 6

2 3 6

1,

1,

1,

1,

1.

max 1 + 2 + 3 + 4 + 5 + 6 M,

i, {0, 1}

1 2 5

1 3

2 4

3 6

2 3 6

1,

1,

1,

1,

1.

max (M+1)1 + (M+1)2 + 3 + 4 + (M+1) 5 + 6 M,

i, {0, 1}

1 2 5

1 3

2 4

3 6

2 3 6

1,

1,

1,

1,

1.

max (M+1)1 + (M+1)2 + 3 + 4 + (M+1) 5 + 6 M,

Page 19: 1 Lecture 5 Set Packing Problems Set Partitioning Problems

19

Further CommentsFurther Comments

set covering problems different from set set covering problems different from set packing problems and set partitioning packing problems and set partitioning problemsproblems

possible to transform a set packing problem possible to transform a set packing problem into a set covering problem, but in general into a set covering problem, but in general not the other way aroundnot the other way around

set covering problem more difficult to solve set covering problem more difficult to solve than the other two problemsthan the other two problems

Page 20: 1 Lecture 5 Set Packing Problems Set Partitioning Problems

20

A Simplified A Simplified Air Crew Scheduling ProblemAir Crew Scheduling Problem

Page 21: 1 Lecture 5 Set Packing Problems Set Partitioning Problems

21

A Simplified A Simplified Air Crew Scheduling ProblemAir Crew Scheduling Problem

six flights every day, for cities six flights every day, for cities AA, , BB, and , and

C, C, where where AA is the base is the base8 10 12 181614 2220

A

B

C

leg 1 leg 2leg 3

leg 4

leg 5 leg 6

Page 22: 1 Lecture 5 Set Packing Problems Set Partitioning Problems

22

A Simplified A Simplified Air Crew Scheduling ProblemAir Crew Scheduling Problem

pairings of legs for air crew: pairings of legs for air crew: rules from rules from regulation bodies and unionregulation bodies and union

simplified rules in the examplesimplified rules in the example at most eight hours flying time in a pairingat most eight hours flying time in a pairing

flying time in a pairing = sum of flying times in all legs of the flying time in a pairing = sum of flying times in all legs of the pairingpairing

at most two duties in a pairingat most two duties in a pairing at least nine hours for overnight rest (OR) between at least nine hours for overnight rest (OR) between

dutiesduties

Page 23: 1 Lecture 5 Set Packing Problems Set Partitioning Problems

23

A Simplified A Simplified Air Crew Scheduling ProblemAir Crew Scheduling Problem

cost of a pairing = time away from the cost of a pairing = time away from the

base base flying time of the pairing flying time of the pairing cost of pairing 1 = 36cost of pairing 1 = 3610 = 2610 = 26

Page 24: 1 Lecture 5 Set Packing Problems Set Partitioning Problems

24

A Simplified A Simplified Air Crew Scheduling ProblemAir Crew Scheduling Problem

suppose only considering covering the 6 suppose only considering covering the 6

legs in a day legs in a day

let let xxjj = 1 if the = 1 if the jjth pairing is used, and th pairing is used, and xxjj = =

0 otherwise0 otherwise

Page 25: 1 Lecture 5 Set Packing Problems Set Partitioning Problems

25

A Simplified A Simplified Air Crew Scheduling ProblemAir Crew Scheduling Problem

a set partitioning problema set partitioning problem min 26min 26xx11 + 20 + 20xx22 + 2 + 2xx33 + 26 + 26xx44 + 20 + 20xx55 + 26 + 26xx66, , ss..tt..

xx11 + + xx22 = 1, = 1,

xx11 + + xx66 = 1, = 1,

xx33 + + xx44 + + xx55 + + xx66 = 1, = 1,

xx22 + + xx33 + + xx44 + + xx55 = 1, = 1,

xx22 + + xx55 = 1, = 1,

xx55 + + xx66 = 1, = 1,

xxii {0, 1} {0, 1}

Page 26: 1 Lecture 5 Set Packing Problems Set Partitioning Problems

26

A Simplified A Simplified Air Crew Scheduling ProblemAir Crew Scheduling Problem

The dual of the partitioning problemThe dual of the partitioning problem min min 11 + + 22 + + 33 + + 44 + + 55 + 26 + 2666,,

ss..tt..

11 + + 22 ≤ 26 ≤ 26

11 + + 44 + + 55 ≤ 20 ≤ 20

33 + + 44 ≤ 2≤ 2

33 + + 44 ≤ 26≤ 26

33 + + 44 + + 55 + + 66 ≤ 20≤ 20

22 + + 33 + + 66 ≤ 26≤ 26

Page 27: 1 Lecture 5 Set Packing Problems Set Partitioning Problems

27

To Construct a SimplifiedTo Construct a SimplifiedAir Crew Scheduling ProblemAir Crew Scheduling Problem

Page 28: 1 Lecture 5 Set Packing Problems Set Partitioning Problems

Bangkok

Singapore

Hong Kong

Taipei

Seoul

3 hr 40min

2 hr 15 min3 hr 45 min

1 hr 30 min

3 hr 10 min2 hr 20 min

28

Six Cities Six Cities

1

4

3

0

2

5

2 hrTokyo

8 am 012 0 flight 9 am 034 5 0 flight

Page 29: 1 Lecture 5 Set Packing Problems Set Partitioning Problems

It takes 90 minutes to load supply and passengers It takes 90 minutes to load supply and passengers

before departure (including unloading passengers for before departure (including unloading passengers for

the previous flight leg, if applicable).the previous flight leg, if applicable).

All flights are on exact time.All flights are on exact time.

It takes 30 minutes for a plane to reach the terminal It takes 30 minutes for a plane to reach the terminal

after arrival.after arrival.

It takes 30 minutes to unload passengers and clean up It takes 30 minutes to unload passengers and clean up

a plane at the end of its service. a plane at the end of its service.

29

Assumptions for PlanesAssumptions for Planes

Page 30: 1 Lecture 5 Set Packing Problems Set Partitioning Problems

30

Itinerary of Itinerary of the Plane for 0-1-2-0 Tourthe Plane for 0-1-2-0 Tour

Time Activities

7:30 Loading

9:00 Leaving Taipei

11:20 Arriving Seoul

11:50 Parked at Terminal; unloading and re-loading passengers; loading suppliers

13:20 Leaving Seoul

15:20 Arriving Tokyo

15:50 Parked at Terminal; unloading and re-loading passengers; loading suppliers

17:20 Leaving Tokyo

20:30 Arriving Taipei

21:00 Parked at Terminal; unloading passengers

21:30 Parking overnight

Page 31: 1 Lecture 5 Set Packing Problems Set Partitioning Problems

31

ExerciseExercise

give the itinerary of the plane for 0-3-give the itinerary of the plane for 0-3-4-5-0 tour4-5-0 tour

based on the itineraries of the planes based on the itineraries of the planes for the 0-1-2-0 and 0-3-4-5-0 tours, for the 0-1-2-0 and 0-3-4-5-0 tours, construct the requirements for air construct the requirements for air stewards and for pilots stewards and for pilots put down all assumptions madeput down all assumptions made