Greedy
EECS 3101
Prof. Andy Mirzaian
Greedy
Algorithms
STUDY MATERIAL:
[CLRS] chapter 16
Lecture Notes 7, 8
2
TOPICS
Combinatorial Optimization Problems
The Greedy Method
Problems:
Coin Change Making
Event Scheduling
Interval Point Cover
More Graph Optimization problems considered later
3
COMBINATORIAL OPTIMIZATION
find the best solution out of finitely many possibilities.
4
Mr. Roboto:
Find the best path from A to B avoiding obstacles
A
B
A
B
There are infinitely many ways to get from A to B.
We can’t try them all.
But you only need to try finitely many critical paths
to find the best.
With brute-force there are exponentially many paths to try.
There may be a simple and fast (incremental) greedy strategy.
5
Mr. Roboto:
Find the best path from A to B avoiding obstacles
The Visibility Graph: 4n + 2 vertices
(n = # rectangular obstacles)
A
B
6
Combinatorial Optimization Problem (COP)
INPUT: Instance I to the COP.
Feasible Set: FEAS(I) = the set of all feasible (or valid) solutions for instance I,
usually expressed by a set of constraints.
Objective Cost Function:
Instance I includes a description of the objective cost function,
CostI that maps each solution S (feasible or not) to a real number or ±.
Goal: Optimize (i.e., minimize or maximize) the objective cost function.
Optimum Set:
OPT(I) = { Sol FEAS(I) | CostI (Sol) CostI (Sol’), Sol’FEAS(I) }
the set of all minimum cost feasible solutions for instance I
Combinatorial means the problem structure implies that only a discrete & finite
number of solutions need to be examined to find the optimum.
OUTPUT: A solution Sol OPT(I), or report that FEAS(I) = .
7
GREEDY METHOD
Greedy attitude:
Don’t worry about the long term consequences of your current action.
Just grab what looks best right now.
8
For such problems it may be possible to build up a solution incrementally
by considering one input object at a time.
GREEDY METHOD is one of the most obvious & simplest such strategies:
Selects the next input object x that is the incremental best option at that time.
Then makes a permanent decision on x (without any backtracking),
either committing to it to be in the final solution, or
permanently rejecting it from any further consideration.
C =
Committed objects
R =
Rejected objects
U =
Undecided objects
S =
Set of
all input
objects
Such a strategy may not work on every problem.
But if it does, it usually gives a very simple & efficient algorithm.
Most COPs restrict a feasible solution to be a finite set or sequence from among the objects described in the input instance.
E.g., a subset of edges of the graph that forms a valid path from A to B.
9
Obstacle avoiding shortest A-B path
A
B
d(p,q) = straight-line distance between points p and q.
u = current vertex of the Visibility Graph we are at (u is initially A).
If (u,B) is a visibility edge, then follow that straight edge. Done.
Otherwise,
from among the visibility edges (u,v) that get you closer to B, i.e., d(v,B) < d(u,B),
make the following greedy choice for v:
(1) Minimize d(u,v) take shortest step while making progress
(2) Minimize d(v,B) get as close to B as possible
(3) Minimize d(u,v) + d(v,B) stay nearly as straight as possible
Which of these greedy choices is guaranteed to produce the shortest A-B path?
u
v
B
greedy choice (1)
greedy choice (2)
greedy choice (3)
shortest path
Later we will investigate a greedy strategy that works.
10
Conflict free object subsets
S = the set of all input objects in instance I
FEAS(I) = the set of all feasible solutions
Definition: A subset C S is “conflict free” if it is contained in some
feasible solution:
true if Sol FEAS(I): C Sol
ConflictFreeI (C) =
false otherwise
ConflictFreeI (C) = false
means C has some internal conflict and no matter what additional input objects you add to it, you cannot get a feasible solution that way.
E.g., in any edge-subset of a simple path from A to B, every vertex has
in-degree at most 1. So a subset of edges, in which some vertex has
in-degree more than 1, has conflict.
11
The Greedy Loop Invariant
The following are equivalent statements of the generic greedy Loop Invariant:
The decisions we have made so far are safe, i.e.,
they are consistent with some optimum solution (if there is any feasible solution).
Either there is no feasible solution, or
there is at least one optimum solution Sol OPT(I) that
includes every object we have committed to (set C), and
excludes every object that we have rejected (set R = S – U – C).
S =
Set of
all input
objects
R
C
U
Sol
LI: FEAS(I) = or [ SolOPT(I) : C Sol C U ].
12
Pre-Cond: S is the set of objects in input instance I
Post-Cond: output COPT(I) or FEAS(I) =
U S
C
LI: FEAS(I) = or
Sol OPT(I) :
C Sol C U
return C
Greedy Choice:
select x U with best
Cost(C{x}) possible
U U – {x} § permanently decide on x
if ConflictFree(C{x}) &
Cost(C{x}) better than Cost(C)
then C C {x} § ----------- commit to x
§ else ------------------------------------ reject x
YES
NO
FEAS(I) = or C OPT(I)
U =
MP = |U|
LI & exit-cond
PostLoopCond
Pre-Cond &
PreLoopCode
LI
PostLoopCond & PostLoopCode
Post-Cond
LI & exit-cond
& LoopCode LI ???
Greedy
method
13
LI & exit-cond & LoopCode LI
U & LI: FEAS(I) = or Sol OPT(I) : C Sol C U
Greedy Choice:
select x U with maximum
Cost(C{x}) possible
U U – {x} § permanently decide on x
if ConflictFree(C{x}) &
Cost(C{x}) > Cost(C)
then C C {x} § ———– commit to x
§ else ———————————— reject x
LI: FEAS(I) = or Solnew OPT(I) : C Solnew C U
Case 1:
x committed
Case 2:
x rejected
NOTE
Solnew
may or may not be the same as Sol
14
14
LI & exit-cond & LoopCode LI
U & LI: FEAS(I) = or Sol OPT(I) : C Sol C U
Greedy Choice:
select x U with maximum
Cost(C{x}) possible
U U – {x} § permanently decide on x
if ConflictFree(C{x}) &
Cost(C{x}) > Cost(C)
then C C {x} § ———– commit to x
§ else ———————————— reject x
LI: FEAS(I) = or Solnew OPT(I) : C Solnew C U
Case 1a:
x committed
and
x Sol
OK. Take
Solnew =Sol
Solnew
may or may not be the same as Sol
15
LI & exit-cond & LoopCode LI
U & LI: FEAS(I) = or Sol OPT(I) : C Sol C U
Greedy Choice:
select x U with maximum
Cost(C{x}) possible
U U – {x} § permanently decide on x
if ConflictFree(C{x}) &
Cost(C{x}) > Cost(C)
then C C {x} § ———– commit to x
§ else ———————————— reject x
LI: FEAS(I) = or Solnew OPT(I) : C Solnew C U
Case 1b:
x committed
and
x Sol
Needs
Investigation
Solnew
may or may not be the same as Sol
16
LI & exit-cond & LoopCode LI
U & LI: FEAS(I) = or Sol OPT(I) : C Sol C U
Greedy Choice:
select x U with maximum
Cost(C{x}) possible
U U – {x} § permanently decide on x
if ConflictFree(C{x}) &
Cost(C{x}) > Cost(C)
then C C {x} § ———– commit to x
§ else ———————————— reject x
LI: FEAS(I) = or Solnew OPT(I) : C Solnew C U
Case 2a:
x rejected
and
x Sol
OK. Take
Solnew =Sol
Solnew
may or may not be the same as Sol
17
LI & exit-cond & LoopCode LI
U & LI: FEAS(I) = or Sol OPT(I) : C Sol C U
Greedy Choice:
select x U with maximum
Cost(C{x}) possible
U U – {x} § permanently decide on x
if ConflictFree(C{x}) &
Cost(C{x}) > Cost(C)
then C C {x} § ———– commit to x
§ else ———————————— reject x
LI: FEAS(I) = or Solnew OPT(I) : C Solnew C U
Case 2b:
x rejected
and
x Sol
Needs
Investigation
Solnew
may or may not be the same as Sol
18
Sol OPT
R
LI & exit-cond & LoopCode LI
Case 1b: x committed and x Sol
Show:
Solnew OPT
C
y
Trade off
x
Show y Sol – C such that
Solnew Sol {x} – {y} satisfies:
(1) Solnew FEAS, &
(2) Cost(Solnew) is no worse than Cost(Sol)
Some times more than one object on either side is traded off.
19
Sol OPT
R
LI & exit-cond & LoopCode LI
Case 2b: x rejected and x Sol
Show:
Solnew OPT
C
y
Trade off
x
Show that
Solnew Sol {y} – {x}, for some y U – Sol
satisfies:
(1) Solnew FEAS, &
(2) Cost(Solnew) is no worse than Cost(Sol)
x could not have created a conflict. So, it must have not improved the objective function.
Some times more than one object on either side is traded off.
20
COIN
CHANGE MAKING
Use minimum number of coins to make change for a given amount.
Greedy: pick the largest coin that fits first, then iterate.
21
The Coin Change Making Problem
Greedy Strategy (the obvious one):
Commit to the largest coin denomination that fits within the (remaining) amount, then iterate.
Greedy(S) = the greedy solution for amount S.
Optimum(S) = the optimum solution for amount S.
PROBLEM: Within a coin denomination system, make change for a given amount S,
using the fewest number of coins possible.
Example 2: Imaginary coin denomination system: 7, 5, 1 kraaps.
S = 10 = 5 + 5 (Optimum) = 7 + 1 + 1 + 1 (Greedy).
Greedy(10) Optimum(10) in this system.
Example 1: Canadian coin denomination system: 25, 10, 5, 1 cents.
S = 98 (Two of many feasible solutions are shown below.)
= 10 + 10 + 10 + 10 + 10 + 10 + 10 + 10 + 10 + 5 + 1 + 1 + 1 13 coins used
= 25 + 25 + 25 + 10 + 10 + 1 + 1 + 1 Optimum sol uses 8 coins.
Greedy(98) = Optimum(98) in the Canadian system.
22
22
The Problem Formulation
INPUT: Coin denomination system a = a1 , a2 , … , an , a1 > a2 > … > an = 1,
and S (all positive integers).
OUTPUT: The solution x = x1 , x2 , … , xn , where
xt = the number of coin type at used, for t = 1..n.
FEAS: a1x1 + a2x2 + … + anxn = S, &
xt is a non-negative integer, for t=1..n.
GOAL: Minimize objective cost x1 + x2 + … + xn.
minimize x1 + x2 + ··· + xn
subject to:
(1) a1x1 + a2x2 + ··· + anxn = S
(2) xtN, for t = 1..n
objective
function
feasibility
constraints
We need an =1 to have a feasible soln for every S.
23
Greedy Choice: choose the largest coin that fits
Conflict Free:
Problem Objective Cost:
Greedy Objective Cost:
At the end: Greedy Objective = Problem objective
( is an unspecified prohibitively large positive number)
The Greedy Choice & Objective
24
The Greedy Loop Invariant
Generic Greedy Loop Invariant ( feasible solution) :
SolOPT : C Sol C U.
Current greedy solution: x1, x2, …, xn
Committed to: x1, x2, …, xt , for some t 1..n.
Rejected: no more of a1, a2, …, at-1.
Considering: any more of at?
Not yet considered: xk = 0 for k > t.
There is U amount remaining to reach the target value S.
Loop Invariant:
x1, x2, …, xn FEAS(S – U), (need U more to reach target S)
Sol = y1, y2, …, yn OPT(S):
yk = xk , for k < t, (Sol excludes what Greedy has rejected)
yt xt , (Sol includes what Greedy has committed to)
xk = 0 , for k > t. (not yet considered)
25
Pre-Cond: input is a = a1, a2, …, an, a1>a2> … >an =1; and S (all pos integers)
Post-Cond: output x1, x2, …, xn OPT(S)
U S; t 1; x1, x2, …, xn 0, 0, …, 0
LI: x1, x2, …, xn FEAS(S – U),
Sol = y1, y2, …, yn OPT(S):
yk = xk for k < t, yt xt , xk =0 for k > t.
return x1, x2, …, xn
if U < at
then t t+1 § reject: no more at
else xt xt +1 § commit to another at
U U - at
YES
NO
U = 0
Algorithm: Greedy Coin Change
LI & exit-cond
PostLoopCond
Pre-Cond &
PreLoopCode
LI
PostLoopCond & PostLoopCode
Post-Cond
LI & exit-cond
& LoopCode LI ???
x1, x2, …, xn OPT(S)
MP = U + (n-t)
26
Efficient implementation
minimize x1 + x2 + ··· + xn
subject to:
(1) a1x1 + a2x2 + ··· + anxn = S
(2) xtN, for t = 1..n
objective
function
feasibility
constraints
Algorithm Greedy( a1 , a2 , … , an , S) § takes O(n) time
CoinCount 0; U S
for t 1 .. n do § largest coin denomination first
xt U div at § xt U/at
U U mod at § U U – at xt
CoinCount CoinCount + xt § objective value
end-for
return ( x1 , x2 , … , xn ; CoinCount)
end
27
Is G(S) = Opt(S) ?
A coin denomination system is called Regular if in that system G(S) = Opt(S) S.
Questions:
(1) In the Canadian Coin System, is G(S) = Opt(S) for all S?
(2) In a given Coin System, is G(S) = Opt(S) for all S?
(3) In a given Coin System, is G(S) = Opt(S) for a given S?
Answers:
(1) YES. It’s Regular. We will see why shortly.
(2) YES/NO: Yes if the system is Regular. NO otherwise.
There is a polynomial time Regularity Testing algorithm
(described next) to find the YES/NO answer.
(3) YES/NO: Regular system: always yes.
Non-regular system: YES for some S, NO for other S.
Exponential time seems to be needed to find the Yes/No answer.
(This is one of the so called NP-hard problems we will study later.)
28
How to Test Regularity
Question: Is a given Coin Denomination System Regular,
i.e., in that system is G(S) = Opt(S) for all S?
Greedy( a1 , a2 , … , an , S) = (x = x1 , x2 , … , xn ; G(S) = St xt ) Optimum( a1 , a2 , … , an , S) = (y = y1 , y2 , … , yn ; Opt(S) = St yt )
YES or NO: For the coin denomination system a1 , a2 , … , an :
S: G(S) = Opt(S).
There are infinitely many values of S to try.
However, if there is any counter-example S, there must be one among
polynomially many critical values that need to be tested to
determine the Yes/No answer. And, each test can be done fast.
What are these critical values? How are they tested?
29
Regularity & Critical Values
Consider the smallest multiple of each coin value that is not smaller than the next larger coin value:
St = at mt, mt = at-1 / at , for t = 2..n.
Necessary Condition for correctness of Greedy:
G(St) mt, for t = 2..n. WHY?
This also turns out to be sufficient (under some mild pre-condition that is satisfied by virtually any reasonable coin denomination system):
FACT 1: [Regularity Theorem of Magazine, Nemhauser, Trotter, 1975]
Pre-Condition: St < at-2 , for t = 3..n
S: G(S) = Opt(S) G(St) mt, for t = 2..n.
Proof: See Lecture Note 7.
30
Let’s Test the Canadian System
t 1 2 3 4 5 6
coin at 200 100 25 10 5 1
mt = at-1 / at 2 4 3 2 5
St = at mt 200 100 30 10 5
Greedy: G(St) 1 1 2 1 1
Pre-Cond: St < at-2 yes yes yes yes
Test: G(St) mt yes yes yes yes yes
This table can be constructed and tested in O(n2) time.
FACT 1: [Magazine, Nemhauser, Trotter, 1975]
Pre-Condition: St < at-2 , for t = 3..n
S: G(S) = Opt(S) G(St) mt, for t = 2..n.
31
Another System
t 1 2 3 4 5 6
coin at 200 100 25 11 5 1
mt = at-1 / at 2 4 3 3 5
St = at mt 200 100 33 15 5
Greedy: G(St) 1 1 5 5 1
Pre-Cond: St < at-2 yes yes yes yes
Test: G(St) mt yes yes NO NO yes
This table can be constructed and tested in O(n2) time.
FACT 1: [Magazine, Nemhauser, Trotter, 1975]
Pre-Condition: St < at-2 , for t = 3..n
S: G(S) = Opt(S) G(St) mt, for t = 2..n.
32
What if Pre-Condition doesn’t hold?
FACT 2: [Pearson, 2005]
S: G(S) = Opt(S) O(n2) critical values test OK in O(n3) time.
FACT 1: [Magazine, Nemhauser, Trotter, 1975]
Pre-Condition: St < at-2 , for t = 3..n
S: G(S) = Opt(S) G(St) mt, for t = 2..n.
[This can be tested in O(n2) time.]
See also Lecture Note 7.
33
The Optimum Sub-Structure Property
We just noticed an important property that will be used many times later:
The optimum sub-structure property: any sub-structure of an optimum structure is itself an optimum structure (for the corresponding sub-instance).
Problems with this property are usually amenable to more efficient algorithmic solutions
than brute-force or exhaustive search methods.
This property is usually shown by a “cut-&-paste” argument (see below).
Example 1: The Coin Change Making Problem.
Consider an optimum solution Sol OPT(S). Let G1 be a group of coins in Sol.
Suppose G1 FEAS(U). Then we must have G1 OPT(U). Why?
Because if G1 OPT(U), then we could cut G1 from Sol, and paste in the optimum sub-structure G2 OPT(U) instead. By doing so, we would get a new solution Sol’ FEAS(S) that has an even better objective value than Sol. But that would contradict Sol OPT(S).
Example 2: The Shortest Path Problem.
Let P be a shortest path from vertex A to vertex B in the given graph G.
Let P’ be any (contiguous) sub-path of P. Suppose P’ goes from vertex C to D.
Then P’ must be a shortest path from C to D. If it were not, then there must be an even shorter path P’’ that goes from C to D. But then, we could replace P’ portion of P by
P’’ and get an even shorter path than P that goes from A to B.
That would contradict the optimality of P.
34
EVENT SCHEDULING
A banquet hall manager has received a list of reservation requests
for the exclusive use of her hall for specified time intervals
She wishes to grant the maximum number of reservation requests
that have no time overlap conflicts
Help her select the maximum number of conflict free time intervals
35
The Problem Statement
INPUT: A set S = { I1, I2, ···, In} of n event time-intervals Ik = sk , fk, k =1..n,
where sk = start time of Ik ,
fk = finishing time of Ik ,
( sk < fk ).
OUTPUT: A maximum cardinality subset C S of mutually compatible intervals
(i.e., no overlapping pairs).
time
Example:
S = the intervals shown below,
C = the blue intervals, C is not the unique optimum.
|C| =4. Can you spot another optimum solution?
36
Some Greedy Heuristics
Greedy iteration step:
From among undecided intervals, select the interval I that looks BEST.
Commit to I if it’s conflict-free
(i.e., doesn’t overlap with the committed ones so far).
Reject I otherwise.
Greedy 1: BEST = earliest start time (min sk).
Greedy 2: BEST = latest finishing time (max fk).
Greedy 3: BEST = shortest interval (min fk – sk).
Greedy 4: BEST = overlaps with fewest # of undecided intervals.
NONE WORKS
37
Earliest Finishing Time First
S = the set of n given intervals
C = committed intervals
R = rejected intervals
U = undecided intervals
MaxF(X) = max{ fk | Ik X }
MinF(X) = min{ fk | Ik X }
Last = MaxF(C)
LOOP INVARIANT:
Sol Opt(S) : C Sol C U,
MaxF(C) = Last MinF(U).
time
1
2
3
4
5
6
7
8
9
Last
Last
Last
Last
C { Ik S | fk Last }
C R { Ik S | fk < Last }
U { Ik S | fk Last }
38
Pre-Cond: input is S = {I1, I2, ···, In} of n intervals, Ik = sk , fk, sk < fk , k =1..n
Post-Cond: output C OPT(S)
U S; C ; Last –
LI: Sol Opt(S):
C Sol C U,
MaxF(C) = Last MinF(U)
return C
Greedy choice: select Ik U with min fk
U U – {Ik } § decide on Ik
if sk Last
then C C {Ik } § commit to Ik
Last fk
§ else --------------------------- reject Ik
YES
NO
U =
Algorithm: Greedy Event Schedule
LI & exit-cond
PostLoopCond
Pre-Cond &
PreLoopCode
LI
PostLoopCond & PostLoopCode
Post-Cond
LI & exit-cond
& LoopCode LI ???
C OPT(S)
MP = |U|
39
LI & exit-cond & LoopCode LI
U &
LI: Sol Opt(S): C Sol C U,
MaxF(C) = Last MinF(U)
Case 1:
Ik rejected
Case 2:
Ik committed
Solnew
may or may not be the same as Sol
Greedy choice: select Ik U with min fk
U U – {Ik } § decide on Ik
if sk Last
then C C {Ik } § commit
Last fk
§ else --------------------------- reject
LI: Solnew Opt(S): C Solnew C U,
MaxF(C) = Last MinF(U)
40
LI & exit-cond & LoopCode LI
U U – {Ik }.
Case 1: sk < Last
Ik rejected
Solnew = Sol maintains LI.
So, C Sol C (U – {Ik }). Thus, Sol still satisfies 1st line of the LI.
Removing Ik from U cannot reduce MinF(U).
Also, C and Last don’t change.
So, 2nd line of the LI is also maintained.
U &
LI: Sol Opt(S): C Sol C U,
MaxF(C) = Last MinF(U)
Last
C
fk
sk
Ik U
sk < Last fk.
Ik has conflict with C,
hence with Sol.
So, Ik Sol.
41
ft
st
It
Sol has no other interval here
LI & exit-cond & LoopCode LI
U &
LI: Sol Opt(S): C Sol C U,
MaxF(C) = Last MinF(U)
U U – {Ik }.
Case 2: sk Last
Ik committed
(C C {Ik }; Last fk)
Solnew = ??? maintains LI.
Last
C
fk
sk
Ik
Ik is conflict-free with C.
So, C{Ik} FEAS(S).
So, | Sol | |C{Ik}| = |C|+1.
So, Sol – C U.
Choose It Sol – C with min ft.
t = k or t k.
It Sol – C U. Greedy choice ft fk.
New solution: Solnew (Sol – {It}) {Ik}.
Solnew is conflict-free & |Solnew| = |Sol| Solnew OPT(S).
& (C {Ik }) Solnew (C {Ik }) (U – {Ik }).
Therefore, Solnew maintains 1st line of the LI.
2nd line of the LI is also maintained.
42
Efficient implementation
Algorithm GreedyEventSchedule( S = I1 , I2 , … , In ) § O(n log n) time
SORT S in ascending order of interval finishing times.
WLOGA I1 , I2 , … , In is the sorted order, i.e., f1 f2 … fn.
C ; Last –
for k 1 .. n do
if Last sk then C C {Ik }
Last fk
end-for
return (C)
end
time
1
2
3
4
5
6
7
8
9
Last
Last
Last
Last
43
INTERVAL
POINT COVER
We have a volunteer group to canvas homes & businesses along Yonge Street.
Each volunteer is willing to canvas a neighboring stretch of houses and shops.
Help us select a minimum number of these volunteers
that can collectively canvas every house and business along the street.
44
The Problem Statement
INPUT: A set P = { p1, p2, … , pn} of n points, and
a set I = { I1= s1 , f1, I2= s2 , f2, … , Im= sm , fm } of m intervals,
all on the real line.
OUTPUT: Find out whether or not I collectively covers P.
If yes, then report a minimum cardinality subset C I of (possibly overlapping) intervals that collectively cover P.
If not, then report a point pP that is not covered by any interval in I.
Example 1:
Example 2:
Not covered by I
45
Some Greedy Heuristics
Greedy choice: pick the incrementally BEST undecided interval first.
Greedy 2: the interval that covers most uncovered points first.
Greedy 1: the longest interval first.
Greedy 3: Let pP be an uncovered point that’s covered by fewest intervals.
If p is not covered by any interval, then report it.
Otherwise, pick an interval that covers p and max # other uncovered points.
NONE WORKS
46
Cover leftmost uncovered point first
C I (committed intervals), U P (uncovered points)
LOOP INVARIANT
(exposed p is not covered by I) &
(I does not cover P or Sol Opt(I,P): C Sol) &
MaxF(C) = Last < Min(U).
Last
Last
Last
Last
GREEDY CHOICE
(1) Pick leftmost point pU (not covered by any interval in C).
If no interval in I covers p, then report that p is exposed.
Else, add to C an interval from I – C that covers p and
(2) extends as far right as possible.
47
Algorithm: Opt Interval Point Cover
Algorithm OptIntervalPointCover ( P = { p1, p2, … , pn}, I= {I1 , I2 , … , Im} )
U P; C ; Last – ; exposed false
while U & not exposed do
p leftmost point in U § greedy choice 1
I’ set of intervals that cover p
if I’ = then exposed true
else do
Select Ik I’ with max fk § greedy choice 2
C C {Ik } ; Last fk
U U – {qP | q Last }
end-else
end-while
if exposed then return ( p is not covered by I )
else return ( C )
end
Loop Invariant: (exposed p is not covered by I) &
(I does not cover P or Sol Opt(I,P): C Sol) &
MaxF(C) = Last < Min(U).
48
Case 1. [exposed = true]:
exposed p is not covered by I. (1st line of LI)
p is not covered by I.
Case 2. [U = & exposed = false]:
C covers P. (3rd line of LI)
C Sol C Opt(I,P). (2nd line of LI)
C is an optimum cover.
PreLoopCode: U P; C ; Last – ; exposed false
exit-cond: U = or exposed.
LI: (exposed p is not covered by I) &
(I does not cover P or Sol Opt(I,P): C Sol) &
MaxF(C) = Last < Min(U).
Lines 1, 2, 3 of the LI become true.
Pre-Cond &
PreLoopCode
LI
LI & exit-cond
PostLoopCond
49
LI & exit-cond & LoopCode LI
Case 1: I’ =
p is exposed
exposed true
LI is maintained.
Suppose It Sol – C covers p. So, It I’. We have t = k or t k.
ft fk (by greedy choice 2).
New solution: Solnew (Sol – {It}) {Ik}.
Solnew covers every point of P covered by Sol , and |Solnew| = |Sol|.
Therefore, Sol OPT(I,P) Solnew OPT(I,P).
C {Ik } Solnew . Therefore, Solnew still maintains 3rd line of the LI.
Remaining lines of LI are also maintained.
U & not exposed &
LI: (exposed p is not covered by I) &
( I does not cover P or
Sol Opt(I,P): C Sol) &
MaxF(C) = Last < Min(U).
Case 2: p covered by Ik I’, max fk
C C {Ik } ; Last fk
U U – {qP | q Last }
Solnew = ??? maintains LI.
Last
C
fk
Ik
p
ft
It
50
Efficient Implementation
To carry out the greedy choices fast:
Line-Scan critical event times t left-to-right on the real line.
Classify each interval Ik = sk , fk I:
Inactive: t < sk ( t hasn’t reached Ik )
Active: sk t fk ( t is covered by Ik )
Dead: fk < t ( t has passed Ik )
Activated.
Classify event e:
e P point event (point activation time = p)
e=(sk , Ik) interval activation event (interval activation time = sk)
ActInts = Max Priority Queue of activated (= active/dead) intervals Ik [priority = fk ].
Events = Min Priority Queue of unprocessed events [priority = activation time].
Iterate: e DeleteMin(Events); process event e.
Minor trick: to avoid DeleteMax on empty ActInts, insert as first activated event a dummy interval I0 = - , -. I0 will remain in ActInts till the end.
51
Algorithm: Efficient implementation
Algorithm OptIntervalPointCover ( P = { p1, p2, … , pn}, I= {I1 , I2 , … , Im} )
C ; Last – ; I0 – , – ; I I {I0}
MakeEmptyMaxHeap(ActInts)
Events ConstructMinHeap(P I) § O(n+m) time
while Events do § O(n + m) iterations
e DeleteMin(Events) § next event to process, O(log(n+m)) time
if e = (sk , Ik) then Insert(Ik , ActInts) § activate interval
else § event e is a point in P
if e > Last then do § greedy choice 1: e = leftmost uncovered point
Ik DeleteMax(ActInts) § greedy choice 2, O(log m) time
if fk < e then return ( point eP is not covered by I )
else do
C C {Ik}
Last fk
end-else
end-if
end-while
return ( C )
end
O( (n+m) log(n+m) ) time
52
Bibliography
If you want to dig deeper, roots of greedy algorithms are in the theory of matroids:
Hassler Whitney, “On the abstract properties of linear dependence,” American Journal of Mathematics, 57:509-533, 1935.
Jack Edmonds, “Matroids and the greedy algorithm,” Mathematical Programming, 1:126-136, 1971.
Eugene L. Lawler, “Combinatorial Optimization: Networks and Matroids,” Holt, Rinehart, and Winston, 1976.
Christos Papadimitriou and Kenneth Steiglitz, “Combinatorial Optimization: Algorithms and Complexity,” Prentice-Hall, 1982. [2nd edition, Courier Dover (publisher), 1998.]
The two cited references on the coin change making problem are:
M.J. Magazine, G.L. Nemhauser, L.E. Trotter, Jr., “When the greedy solution solves a class of knapsack problems,” Operations Research, 23(2):207-217, 1975.
David Pearson “A polynomial-time algorithm for the change-making problem,” Operations Research Letters, 33(3):231-234, 2005.
53
Exercises
54
The shortest obstacle avoiding path: As we discussed, the scene consists of a pair of points A and B among n pairwise disjoint rectangular obstacles with horizontal and vertical sides. We listed 3 greedy heuristics. We saw that all three fail to find the shortest obstacle avoiding A-B path on some instances.
An interesting question arises: How badly can these heuristics fail?
(a) Explore to find the worst scene for each of these heuristics. By worse, we mean the
ratio of the length of the path found by the heuristic compared with the length of
the shortest path, expressed as a function of n, is as high as possible. How bad
could it be? Is it unbounded? Supper-linear? Linear? Sub-linear? …
(b) Does the answer improve if the obstacles are congruent squares?
Coin Change making: For each of the following coin denomination systems
either argue that the greedy algorithm always yields an optimum solution for
any given amount, or give a counter-example:
(a) Coins c0 , c1 , c2 , …, cn-1, where c is an integer > 1.
(b) Coins 1, 7, 13, 19, 61.
(c) Coins 1, 7, 14, 20, 61.
[CLRS, Exercise 16.2-5, page 428] Smallest unit length interval covering set:
Describe an efficient algorithm that, given a set P = { p1, p2, … , pn} of points on the real line, determines the smallest set of unit-length closed intervals that covers all of the given points. Argue that your algorithm is correct.
Interval Point Cover: What is the time complexity of the Interval Point Cover problem? We developed an O((n+m) log (n+m)) time algorithm for this problem.
Derive a lower bound on this problem by a reduction from the Min-Gap Problem.
55
[CLRS, Exercise 16.1-4, page 379] Minimum number of lecture halls:
Suppose we have a set of events to schedule among a large number of lecture halls.
We wish to schedule all the events using as few lecture halls as possible.
Design and analyze an efficient greedy algorithm to determine which event should use which lecture hall. Prove the optimality of the solution produced by your algorithm.
[This is also known as the interval-graph colouring problem. We can create an interval graph whose vertices are the given events and whose edges connect incompatible events. The smallest number of colours required to colour every vertex
so that no two adjacent vertices are given the same colour corresponds to finding the fewest lecture halls needed to schedule all the given events.]
Smallest Hitting Set: Design and analyze an efficient greedy algorithm for the following problem:
Input: A set P = { p1, p2, … , pn} of points, and a set I = { I1, I2, … , Im} of intervals, all on the real line. These intervals and points are given in no particular order. Each interval is given by its starting and finishing times.
Output: (i) A minimum cardinality subset H of P such that every interval in I is hit by (i.e., contains) at least one point in H, or
(ii) an interval Ik I that is not hit by any point in P.
Given a set of black and white intervals, select a smallest number of white intervals that collectively overlap every black interval. State your greedy choice and prove its correctness.
56
One Machine Scheduling with Deadlines:
You are given a set {J1, J2, … , Jn} of n jobs to be processed on a single sequential machine. Associated with each job Jk is a processing time tk and a deadline dk by which it must be completed. A feasible schedule is a permutation of the jobs such that if the jobs are processed in that order, then each job finishes by its deadline.
Design & analyze a simple greedy strategy that finds a feasible schedule if there is any.
Two Machine Scheduling:
We are given a set {J1, J2, … , Jn} of n jobs that need to be processed by two machines A and B. These machines perform different operations and each can process only one job at a time. Each job has to be processed by both machines; first by A, then by B.
Job Jk requires a given duration Ak on machine A, and a given duration Bk on B.
We wish to find the minimum total duration required to process all n jobs by both machines, as well as the corresponding optimum schedule.
A schedule is the sequencing of the n jobs through the two machines.
Both machines will process the jobs based on the scheduled order. Each job J, in the scheduled order, is first processed on machine A as soon as A completes its previously scheduled job. Upon completion by A, job J is processed by B as soon as B becomes available.
Design and analyze an efficient greedy algorithm for this problem.
Prove the optimality of the schedule produced by your algorithm.
57
The widely popular Spanish search engine El Goog needs to do a serious amount of computation every time it recompiles its index. Fortunately, the company has at its disposal a single large supercomputer, together with an essentially unlimited supply of high-end PCs.
They have broken the overall computation into n distinct jobs, labeled J1, J2, …, Jn, which can be performed completely independently of one another. Each job consists of two stages: first it needs to be preprocessed on the supercomputer, and then it needs to be finished on one of the PCs. Let’s say that job Jk needs pk seconds of time on the supercomputer, followed by fk seconds of time on a PC.
Since there are at least n PCs available on the premises, the finishing of the jobs can be performed fully in parallel – all the jobs can be processed at the same time. However, the supercomputer can only work on a single job at a time. So the system managers need to work out an order in which to feed the jobs to the supercomputer. As soon as the first job in order is done on the supercomputer, it can be handed off to a PC for finishing; at that point in time a second job can be fed to the supercomputer; when the second job is done on the supercomputer, it can proceed to a PC regardless of whether or not the first job is done (since PCs work independently in parallel), and so on.
Let’s say that a schedule is an ordering of the jobs for the supercomputer, and the completion time of the schedule is the earliest time at which all jobs will have finished processing on the PCs. This is an important quantity to minimize, since it determines how rapidly El Goog can generate a new index.
Design and analyze an efficient greedy algorithm to find a minimum completion time schedule.
58
The Factory Fueling Problem: A remotely-located factory has a fuel reservoir which, when full, has enough fuel to keep the factory’s machines running for M days. Due to the factory’s remote location, the fuel supply company does not deliver fuel on-demand. Instead, its trucks make visits to the factory’s area according to a preset annual schedule and if requested, would fill the reservoir. The schedule is given as an array S[1..n] where S[i] is the date of the i-th visit in the year. Each time the reservoir is filled, a fixed delivery charge is applied regardless of how much fuel is supplied. The factory management would like to minimize these delivery charges and, at the same time, minimize the idle, empty reservoir periods. Given M and S, describe an efficient greedy algorithm to obtain an optimal annual filling schedule; i.e., determine for each of the n visits whether the reservoir should be filled. Prove that your algorithm satisfies the greedy-choice property.
The Loading Problem: Consider a train that travels from station 1 to station n with intermediate stops at stations 2, 3, …, (n – 1). We have p[i,j] packages that need to be delivered from point i to point j where 1 ≤ i < j ≤ n. Packages have the same size. The maximum number at any point in the train must not exceed its capacity C. We want to deliver as many packages as possible.
a) Give a strategy for dropping and picking up packages at each point.
b) Prove your strategy maximizes the number of delivered packages.
[Hint: Use the greedy idea twice in the solution of this problem. At point 1 load packages in increasing order of their destination till capacity is reached. When the train arrives at point 2, (conceptually) unload and reload according to the same principle as above. If some packages that were picked up at point 1 get left at point 2, do not load them at point 1 in the first place!]
59
Compatible Capacitated Assignment:
Input: Two sorted arrays A[1..n] and B[1..n] of reals; a positive integer capacity C, and a
positive Compatibility real number a.
Definitions:
(i) A pair (i,j) is called compatible, if | A[i] - B[j] | ≤ a.
(ii) An assignment is a pairing among elements of arrays A and B so that each element of each array appears in at least one pairing with an element of the other array.
(iii) An assignment is compatible if each pairing in that assignment is compatible.
(iv) A valid assignment is a compatible one in which no element is paired with more than C elements from the other array.
Output: A valid assignment, or nil if no such assignment exists.
Design analyze and carefully prove the correctness of an efficient greedy algorithm for this
problem. [Hint: Show the following claims are true:
(i) If (i,j1) and (i,j2) are compatible, then so is every pair (i,j) such that j is between j1 and j2. Similarly, if (i1,j) and (i2,j) are compatible, then so is every pair (i,j) such that i is between i1 and i2.
(ii) If there is a valid assignment, then there is a valid assignment with no crossing, where a crossing consists of two assigned pairs (i1,j1) and (i2,j2) such that i1 < i2 but j1 > j2, or
i1 > i2 but j1 < j2.
(iii) An uncrossed valid assignment has the property that if (i,j1) and (i,j2) are pairs in that assignment, then so are all pairs of the form (i,j) with j between j1 and j2. Similarly, if (i1,j) and (i2,j) are pairs in the assignment, then so are all pairs (i,j) such that i is between i1 and i2.
(iv) We can assign pairs by scanning arrays A and B in a greedy merge fashion.]
60
Egyptian Fractions:
Input: A rational fraction x/y, where x and y are integers and 0