1 / 51

Greedy Algorithms

Greedy Algorithms. Greedy Algorithms. Compared to Dynamic Programming Both used in optimization problems More efficient Not always optimal Top-down instead of bottom up Make a locally optimal (quick) choice Still have optimal substructure No need to dive into sub-problems.

bailey
Download Presentation

Greedy Algorithms

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Greedy Algorithms Jeff Chastine

  2. Greedy Algorithms • Compared to Dynamic Programming • Both used in optimization problems • More efficient • Not always optimal • Top-down instead of bottom up • Make a locally optimal (quick) choice • Still have optimal substructure • No need to dive into sub-problems Jeff Chastine

  3. Activity Selection • Given a set of activities , find the maximum # you can schedule • NOT optimal use of time of a room • One activity is active at one time • Activity has start time of and finish of • Order activities by finish time isifi 1 2 3 4 5 6 7 8 9 10 11 1 3 0 5 3 5 6 8 8 2 12 4 5 6 7 8 9 10 11 12 13 14 Jeff Chastine

  4. Using Dynamic Programming • Defining the subproblem: • Set of activities that “fit” in this time chunk • Picking any activity creates two sub-problems Jeff Chastine

  5. Visualization Sij aa ac ak ax az ab aw ay Jeff Chastine

  6. Visualization Skj Sik Sij • Optimal solution includes aa ac ak ax az ab aw ay Jeff Chastine

  7. Visualization Skj Sik Sij • We know • How to choose activity ? aa ac ak ax az ab aw ay Jeff Chastine

  8. Visualization Skj Sik Sij • Iterate through all (like Matrix Mults)! aa ac ak ax az ab aw ay Jeff Chastine

  9. Recursive Solution RECURSIVE-ACT-SELECT (s, f, i, n) 1 m ← i + 1 2 whilem ≤ n and sm < fi //find 1st activity 3 dom ← m + 1 4 ifm ≤ n 5 then return {am} RECURSIVE-ACT-SELECT(s, f, m, n) 6 else return Jeff Chastine

  10. Code Trace a0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Jeff Chastine

  11. First Call a0 a1 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Jeff Chastine

  12. Second Call a2 a0 a1 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Jeff Chastine

  13. Second Call a2 a0 a1 a3 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Jeff Chastine

  14. Second Call a2 a0 a1 a4 a3 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Jeff Chastine

  15. Third Call a0 a1 a4 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Jeff Chastine

  16. Third Call a5 a0 a1 a4 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Jeff Chastine

  17. Third Call a5 a0 a1 a4 a6 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Jeff Chastine

  18. Third Call a5 a0 a1 a4 a6 a7 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Jeff Chastine

  19. Third Call a5 a0 a1 a4 a8 a6 a7 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Jeff Chastine

  20. Fourth Call a0 a1 a4 a8 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Jeff Chastine

  21. Fourth Call a9 a0 a1 a4 a8 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Jeff Chastine

  22. Fourth Call a9 a0 a1 a4 a8 a10 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Jeff Chastine

  23. Fourth Call a9 a0 a1 a4 a8 a11 a10 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Jeff Chastine

  24. Result a0 a1 a4 a8 a11 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Jeff Chastine

  25. Where Greedy Fails • Knapsack Problem • Steal items of different weight and value • Sack has weight limit • Goal: steal as much as possible • Two flavors of problem • 0-1 version (take all of something) • Fractional (can take only part) Jeff Chastine

  26. 0-1 Example 50 item 3 30 item 2 20 item 1 10 $100 $60 $120 Jeff Chastine

  27. Greedy – Go for the Gold! 50 item 3 30 item 2 20 item 1 10 $100 $60 $120 Jeff Chastine

  28. Greedy – Go for the Gold! 50 item 3 30 item 2 20 item 1 10 10 $100 $60 $120 Bad Idea… Jeff Chastine

  29. Greedy – Pick the next best 50 item 3 30 20 item 2 20 item 1 10 10 $100 $60 $120 Total = $160 Jeff Chastine

  30. Greedy Fail 50 30 item 3 30 item 2 20 item 1 10 10 $100 $60 $120 Try again! Total = $180 Jeff Chastine

  31. Optimal for 0-1 Problem 50 30 item 3 30 item 2 20 20 item 1 10 $100 $60 $120 Total = $220 Jeff Chastine

  32. Fractional Example 50 item 3 30 item 2 20 item 1 10 $100 $60 $120 Jeff Chastine

  33. Greedy – Go for the Gold! 50 item 3 30 item 2 20 item 1 10 $100 $60 $120 Jeff Chastine

  34. Greedy – Go for the Gold! 50 item 3 30 item 2 20 item 1 10 10 $100 $60 $120 Jeff Chastine

  35. Greedy – Get next best… 50 item 3 30 20 item 2 20 item 1 10 10 $100 $60 $120 Jeff Chastine

  36. Greedy – AHA! Take only part 50 2030 item 3 30 20 item 2 20 item 1 10 10 $100 $60 $120 Jeff Chastine

  37. Greedy – Go for the Gold! 50 2030 $80 item 3 30 20 item 2 $100 20 item 1 10 10 $60 $100 $60 $120 Total = $240 Note: optimal for 0-1 was $220 Jeff Chastine

  38. Why the Greedy Fail? 50 50 30 item 3 30 20 item 2 20 item 1 10 10 10 $100 $60 $120 Notice the waste of space! Jeff Chastine

  39. Huffman Codes • Objective: compress text data • Raw ASCII text file uses 8 bits/char (fixed) • 1000 characters = 8K bits • Book uses 6 characters for 3 bits/char • Trick • Analyze: not all characters are used! • Analyze: frequency of characters • Prefix codes: no binary code is prefix of another Jeff Chastine

  40. Analysis a b c d e f Frequency (in thousands) 45 13 12 16 9 5 Fixed-length codeword 000 001 010 011 100 101 Variable-length codeword 0 101 100 111 1101 1100 110001001101 = face //12 bits 01011010 = abba // 8 bits Jeff Chastine

  41. How it works • Sort by least frequent e:9 c:12 b:13 d:16 a:45 f:5 Jeff Chastine

  42. How it works • Group lowest two e:9 c:12 b:13 d:16 a:45 f:5 Jeff Chastine

  43. How it works • Combine and re-sort 14 c:12 b:13 d:16 a:45 1 0 e:9 f:5 Watch how least frequent characters are pushed down the tree! Jeff Chastine

  44. How it works • Group lowest two 14 c:12 b:13 d:16 a:45 1 0 e:9 f:5 Jeff Chastine

  45. How it works • Re-sort 14 25 d:16 a:45 1 1 0 0 e:9 f:5 c:12 b:13 Jeff Chastine

  46. How it works 14 25 d:16 a:45 1 1 0 0 e:9 f:5 c:12 b:13 Jeff Chastine

  47. How it works 25 30 a:45 0 1 1 0 14 c:12 b:13 d:16 1 0 e:9 f:5 Notice what’s happening to the first grouping we did Jeff Chastine

  48. How it works 25 30 a:45 0 1 1 0 14 c:12 b:13 d:16 1 0 e:9 f:5 Jeff Chastine

  49. How it works 55 a:45 0 1 25 30 0 1 1 0 14 c:12 b:13 d:16 1 0 e:9 f:5 Jeff Chastine

  50. How it works 100 1 0 55 a:45 0 1 25 30 0 1 1 0 14 c:12 b:13 d:16 1 0 e:9 f:5 Jeff Chastine

More Related