Greedy Algorithms. Kleinberg and Tardos, Chapter 4

Transcription

1 Greedy Algorithms Kleinberg and Tardos, Chapter 4 1

2 Selecting breakpoints Road trip from Fort Collins to Durango on a given route. Fuel capacity = C. Goal: makes as few refueling stops as possible. Greedy algorithm. Go as far as you can before refueling. C C C C Fort Collins C C C Durango 2

3 Selecting Breakpoints: Greedy Algorithm The road trip algorithm. Sort breakpoints so that: 0 = b 0 < b 1 < b 2 <... < b n = L S {0} x 0 breakpoints selected current distance while (x b n ) let p be largest integer such that b p x + C if (b p = x) return "no solution" x b p S S {p} return S 3

4 Coin Changing Goal. Given currency denominations: 1, 5, 10, 25, 100, devise a method to pay amount to customer using fewest number of coins. Example: 34. Cashier's algorithm. At each iteration, add coin of the largest value that does not take us past the amount to be paid. Example: $

5 Coin-Changing: Greedy Algorithm Cashier's algorithm. Use the maximal number of the largest denomination x amount to be changed Sort coins denominations by value: c 1 < c 2 < < c n. S φ coins selected while (x > 0) { let k be largest integer such that c k x if (k = 0) return "no solution found" x x - c k S S {k} } return S Can be proven to be optimal! 5

6 Coin-Changing: Greedy doesn t always work Greedy algorithm works for US coins. Fails when changing 40 when the denominations are 1, 5, 10, 20, 25. 6

7 Interval Scheduling Job j starts at s j and finishes at f j. Two jobs compatible if they don't overlap. Goal: find maximum size subset of compatible jobs. a b c d e f g h Time 7

8 Interval Scheduling: Greedy Algorithms Greedy template. Consider jobs in some natural order. Take each job provided it's compatible with the ones already taken. [Earliest start time] Consider jobs in ascending order of s j. [Earliest finish time] Consider jobs in ascending order of f j. [Shortest interval] Consider jobs in ascending order of f j s j. [Fewest conflicts] For each job j, count the number of conflicting jobs c j. Schedule in ascending order of c j. 8

9 Interval Scheduling: Greedy Algorithms Greedy template. Consider jobs in some natural order. Take each job provided it's compatible with the ones already taken. counterexample for earliest start time counterexample for shortest interval counterexample for fewest conflicts 9

10 Interval Scheduling: Greedy Algorithm Greedy algorithm. Consider jobs in increasing order of finish time. Take each job provided it's compatible with the ones already taken. Sort jobs by finish times so that f 1 f 2... f n. set of jobs selected A φ for j = 1 to n { if (job j compatible with A) A A {j} } return A Implementation. O(n log n). Remember job j* that was added last to A. Job j is compatible with A if s j f j*. 10

11 Interval Scheduling: Analysis Theorem. The greedy algorithm is optimal. Proof. (by contradiction) Assume greedy is not optimal, and let's see what happens. Let i 1, i 2,... i k be the jobs selected by greedy. Let j 1, j 2,... j m be the jobs in the optimal solution with i 1 = j 1, i 2 = j 2,..., i r = j r for the largest possible value of r. job i r+1 finishes before j r+1 Greedy: i 1 i 2 i r i r+1... OPT: j 1 j 2 j r j r+1... why not replace job j r+1 with job i r+1? 11

12 Interval Scheduling: Analysis Theorem. The greedy algorithm is optimal. Proof. (by contradiction) Assume greedy is not optimal, and let's see what happens. Let i 1, i 2,... i k be the jobs selected by greedy. Let j 1, j 2,... j m be the jobs in the optimal solution with i 1 = j 1, i 2 = j 2,..., i r = j r for the largest possible value of r. Greedy: i 1 i 2 i r i r+1... OPT: j 1 j 2 j r i r+1... solution still feasible and optimal 12

13 Interval Scheduling: Analysis Theorem. The greedy algorithm is optimal. Proof. (by contradiction) Assume greedy is not optimal, and let's see what happens. Let i 1, i 2,... i k be the jobs selected by greedy. Let j 1, j 2,... j m be the jobs in the optimal solution with i 1 = j 1, i 2 = j 2,..., i r = j r for the largest possible value of r. By assumption, the optimal solution contains a larger number of jobs. Greedy: i 1 i 2 i r i r+1 OPT: j 1 j 2 j r i r+1... the one with the earliest end time should have been chosen by the greedy algorithm - contradiction 13

14 Greedy algorithms Greedy algorithms determine a globally optimum solution by a series of locally optimal choices

15 Interval Scheduling Lecture j starts at s j and finishes at f j. Goal: find minimum number of classrooms to schedule all lectures so that no two occur at the same time in the same room. This schedule uses 4 classrooms to schedule 10 lectures: 4 e j 3 c d g 2 b h 1 a f i 9 9: : : :30 1 1:30 2 2:30 3 3:30 4 4:30 Time 15

16 Interval Scheduling Lecture j starts at s j and finishes at f j. Goal: find minimum number of classrooms to schedule all lectures so that no two occur at the same time in the same room. This schedule uses 3: 3 c d f j 2 b g i 1 a e h 9 9: : : :30 1 1:30 2 2:30 3 3:30 4 4:30 Time 16

17 Interval Scheduling: Lower Bound Key observation. Number of classrooms needed depth (maximum number of intervals at a time point) Example: Depth of schedule below = 3 schedule is optimal. Q. Does there always exist a schedule equal to depth of intervals? 3 c d f j 2 b g i 1 a e h 9 9: : : :30 1 1:30 2 2:30 3 3:30 4 4:30 Time 17

18 Interval Scheduling: Greedy Algorithm Greedy algorithm. Consider lectures in increasing order of start time and assign lecture to any compatible classroom. allocate d labels(d = depth) sort the intervals by starting time: I 1,I 2,..,I n for j = 1 to n for each interval I i that precedes and overlaps with I j exclude its label for I j pick a remaining label for I j 18

19 Greedy works allocate d labels (d = depth) sort the intervals by starting time: I 1,I 2,..,I n for j = 1 to n for each interval I i that precedes and overlaps with I j exclude its label for I j pick a remaining label for I j Observations: There is always a label for I j No overlapping intervals get the same label

20 Scheduling to Minimize Lateness Minimizing lateness problem. Single resource processes one job at a time. Job j requires t j units of processing time and is due at time d j. If j starts at time s j, it finishes at time f j = s j + t j. Lateness: j = max { 0, f j - d j }. Goal: schedule all jobs to minimize maximum lateness - max j. Example: 1 t j d j lateness = 2 lateness = 0 max lateness = 6 d 3 = 9 d 2 = 8 d 6 = 15 d 1 = 6 d 5 = 14 d 4 =

21 Minimizing Lateness: Greedy Strategies Greedy template. Consider jobs in some order. [Shortest processing time first] Consider jobs in ascending order of processing time t j. [Earliest deadline first] Consider jobs in ascending order of deadline d j. [Smallest slack] Consider jobs in ascending order of slack d j - t j. 21

22 Minimizing Lateness: Greedy Strategies Greedy template. Consider jobs in some order. [Shortest processing time first] Consider jobs in ascending order of processing time t j. d j 1 t j counterexample [Smallest slack] Consider jobs in ascending order of slack d j - t j. 1 t j 1 2 d j counterexample 22

23 Minimizing Lateness: Greedy Algorithm Greedy algorithm. Earliest deadline first. sort jobs by deadline so that d 1 d 2 d n t 0 for j = 1 to n assign job j to interval [t, t + t j ] t t + t j max lateness = 1 d 1 = 6 d 2 = 8 d 3 = 9 d 4 = 9 d 5 = 14 d 6 =

24 Minimizing Lateness: No Idle Time Observation. There exists an optimal schedule with no idle time. d = 4 d = d = d = 4 d = 6 d = Observation. The greedy schedule has no idle time. 24

25 Minimizing Lateness: Inversions Def. Given a schedule A, an inversion is a pair of jobs i and j such that: d i < d j but j scheduled before i. inversion j i Observation. Greedy schedule has no inversions. 25

26 Minimizing Lateness: Inversions Def. Given a schedule A, an inversion is a pair of jobs i and j such that: d i < d j but j scheduled before i. inversion j i Observation. All schedules with no inversions and no idle time have the same maximum lateness How are they different? In the order of jobs with the same deadline are scheduled 26

27 Minimizing Lateness: Inversions Def. Given a schedule A, an inversion is a pair of jobs i and j such that: d i < d j but j scheduled before i. inversion j i Observation. All schedules with no inversions and no idle time have the same maximum lateness If we show that there is an optimal schedule with no inversions the greedy schedule is optimal. 27

28 Minimizing Lateness: Inversions Def. Given a schedule A, an inversion is a pair of jobs i and j such that: d i < d j but j scheduled before i. inversion j i Observation. If a schedule (with no idle time) has an inversion, it has one with a pair of inverted jobs scheduled consecutively. 28

29 Minimizing Lateness: Inversions inversion before swap j i f i after swap i j Claim. Swapping two consecutive, inverted jobs reduces the number of inversions by one and does not increase the max lateness. Proof. Let be the lateness before the swap, and let ' be the lateness afterwards. ' k = k for all k i, j!" ' i j = f j " # d j (definition) i f' j = f i # d j ( j finishes at time f i ) $ f i # d i (i < j) $! i (definition) 29

30 Minimizing Lateness: Analysis of Greedy Algorithm Theorem. Greedy schedule S is optimal. Proof. Let S* to be an optimal schedule that has the fewest number of inversions. Can assume S* has no idle time. If S* has no inversions, then l(s) = l(s*). If S* has an inversion, let i-j be an adjacent inversion. swapping i and j does not increase the maximum lateness and strictly decreases the number of inversions continue until no inversions left At each step lateness not increased, i.e l(s) l(s*), and since S* is optimal, l(s) = l(s*) 30

31 A wordle a word collage. Wordle A wordle constructed out of one of the instructor s papers: wordle constructed using the java applet at wordle.net wordle uses a randomized greedy algorithm to solve the packing problem words are placed randomly and ordered by frequency 31

32 Greedy Analysis Strategies Greedy algorithm stays ahead. Show that after each step of the greedy algorithm, its solution is at least as good as any other. Structural. Discover a simple "structural" bound asserting that every possible solution must have a certain value. Then show that your algorithm always achieves this bound. Exchange argument. Incrementally transform any solution to the greedy one without hurting its quality. Other greedy algorithms. Kruskal, Prim, Dijkstra, Huffman, 32

33 Selecting Breakpoints: Correctness Theorem. Greedy algorithm is optimal. Pf. (by contradiction) Assume greedy is not optimal, and let's see what happens. Let 0 = g 0 < g 1 <... < g p = L denote set of breakpoints chosen by greedy. Let 0 = f 0 < f 1 <... < f q = L denote set of breakpoints in an optimal solution with f 0 = g 0, f 1 = g 1,..., f r = g r for largest possible value of r. Note: g r+1 > f r+1 by greedy choice of algorithm. Greedy: g 0 g 1 g 2 g r g r+1 OPT:... f 0 f 1 f 2 f r f r+1 f q why doesn't optimal solution drive a little further? 33

34 Selecting Breakpoints: Correctness Theorem. Greedy algorithm is optimal. Pf. (by contradiction) Assume greedy is not optimal, and let's see what happens. Let 0 = g 0 < g 1 <... < g p = L denote set of breakpoints chosen by greedy. Let 0 = f 0 < f 1 <... < f q = L denote set of breakpoints in an optimal solution with f 0 = g 0, f 1 = g 1,..., f r = g r for largest possible value of r. Note: g r+1 > f r+1 by greedy choice of algorithm. Greedy: g 0 g 1 g 2 g r g r+1 OPT:... f 0 f 1 f 2 f r f q another optimal solution has one more breakpoint in common contradiction 34