1. Welcome and Introduction 15-211 Fundamental Data Structures and Algorithms Margaret Reid-Miller and Aleks Nanevski January 13, 2004

2. What is this course about?

3.  How to solve computing problems.  Problem analysis, to abstract away details and divide into smaller subproblems.  Mathematical foundations for precise formulations of problems and solutions.  Data structures and algorithms to solve problems correctly and efficiently.  Java programming and modular software construction for good implementations.

4. An Example

5. Generating mazes  We want to write a program to generate mazes.  We want every maze to be solvable.

6. Generating mazes, cont’d  Also: We want mazes to be fun, i.e.,  We want maze solutions to be unique.  We want every “room” to be reachable. How should we think about this?

7. Let’s hack!

8. 2am in the WeH cluster…

9. No, let’s think!

10. Thinking about the problem  Think about a grid of rooms separated by walls.  Each room can be given a name. abcd hgfe ijkl ponm Randomly knock out walls until we get a good maze.

11. Mathematical formulation  A set of rooms:  {a, b, c, d, e, f, g, h, i, j, k, l, m, n, o, p}  Pairs of adjacent rooms that have an open wall between them.  For example, (a,b) and (g,k) are pairs. abcd hgfe ijkl ponm

12. Graphs  Abstractly, this is a mathematical structure called a graph.  Informally, a graph consists of a set of nodes and a set of edges.

13. Mazes as graphs  A maze is a graph where  each node represents a room, and  an edge from node x to node y indicates that rooms x and y are adjacent and there is no wall in between them.  A part of this course will be about graph algorithms.

14. Why math?  Q: Why is it useful to formulate the problem so that mazes are graphs?  A: Data structures are typically defined as mathematical structures.  A: Mathematics can be used to reason about the correctness and efficiency of data structures and algorithms.  A: Mathematical structures make it easier to think — to abstract away from unnecessary details and avoid “hacking”.

15. Mazes as graphs abcd hgfe ijkl ponm {(a,b), (b,c), (a,e), (e,i), (i,j), (f,j), (f,g), (g,h), (d,h), (g,k), (m,n), (n,o), (k,o), (o,p), (l,p)}

16. Mazes as graphs {(a,b), (b,c), (a,e), (e,i), (i,j), (f,j), (f,g), (g,h), (d,h), (g,k), (m,n), (n,o), (k,o), (o,p), (l,p)} abcd efgh ijkl mnop

17. Unique solutions  What property must the graph have for the maze to have a solution?  A path from (a) to (p).  What property must it have for the maze to have a unique solution?  Graph must be a tree. abcd efgh ijkl mnop

18. Mazes as trees  Informally, a tree is a graph:  with a unique root node, and  each node having a unique parent.  A spanning tree is a tree that includes all of the nodes.  Why is it good to have a spanning tree? a b c d e f g h i j k lm n o p Trees have no cycles!

19. What is this course about? 1. Thinking about problems abstractly.

20. Algorithm  Now that we have a data structure in mind, we can think about the algorithm.  Essentially:  Randomly pick a wall and delete it (add it to the tree) if it won’t create a cycle.  Stop when a spanning tree has been created.  This is Kruskal’s Algorithm.

21. Creating a spanning tree  When adding a wall to the tree, how do we detect that it won’t create a cycle?  When adding wall (x,y), we want to know if there is already a path from x to y in the tree.  In fact, there is a fast algorithm for doing exactly this, called union-find.

22. Using the union-find algorithm  We put rooms into an equivalence class if there is a path connecting them.  Before adding an edge (x,y) to the tree, make sure that x and y are not in the same equivalence class. abcd efgh ijkl mnop Partially- constructed maze

23. How fast?  Is this a fast way to generate mazes?  How much time will it take to generate a maze?  What do we mean by “fast”?  In addition to finding the right algorithms, analyzing the performance of algorithms will be a major part of this course.

24. Performance and Scaling  Suppose we have three algorithms to choose from.  Which one to select?  Systematic analysis reveals performance characteristics.  For a problem of size n (i.e., detecting cycles out of n nodes). 1.100n sec 2.7n 2 sec 3.2 n sec

25. Performance and Scaling n 100n sec7n 2 sec2 n sec 1 100 s7 s2 s 5.5 ms 175 s32 s 101 ms.7 ms1 ms 454.5 ms14 ms1 year 100……… 1,000 10,000 1,000,000

26. Performance and Scaling n 100n sec7n 2 sec2 n sec 1 100 s7 s2 s 5.5 ms 175 s32 s 101 ms.7 ms1 ms 454.5 ms14 ms1 year 100……… 1,000 10,000 1,000,000

27. What?! One year?  2 10 = 1,024  1024 sec  1 ms  2 45 = 35,184,372,088,832  3.5 * 10 13 sec =3.5 * 10 7 sec  1.1 year

28. Performance and Scaling n 100n sec7n 2 sec2 n sec 1 100 s7 s2 s 5.5 ms 175 s32 s 101 ms.7 ms1 ms 454.5 ms14 ms1 year 100100 ms7 sec10 16 year 1,0001 sec12 min-- 10,00010 sec20 hr-- 1,000,0001.6 min.22 year--

29. What is this course about? 2. Selecting good data structures and fast algorithms.

30. Modular design  By thinking about the problem, we have strong hints about the structure of our program  Grids, Graphs (with edges and nodes), Spanning trees, Union-find.  With disciplined programming, we can write our program to reflect this structure.  Modular programs are usually easier to get right and easier to understand.

31. Modular design SpanningTree Maze GraphGrid MazeDraw UnionFind Reusable components

32. Is it correct?  How will we know if we implemented our solution correctly?  What do we mean by “correct”?  Will it generate the right answers?  Will it terminate?  We will spend some time in this course on techniques for proving the correctness of programs.

33. What is this course about? 3. Implementing programs that are understandable and correct.

34. Instant Quiz

35. Instant quiz: Two questions  Given that we are using trees to represent mazes, what would be a simple algorithm for solving mazes?  Besides insisting on spanning trees, what other criteria might one want in order to improve the “quality” or “fun factor” of the mazes? a b c d e f g h i j k lm n o p

36. What is this course about?

37. The science in CS  Not “hacking”, but:  Thinking about problems abstractly.  Selecting good data structures and obtaining correct and fast algorithms.  Implementing programs that are understandable and correct.

38. How the Course Works

39. Course staff  A complete listing of the course staff (including instructors and teaching assistants) will be available on the course Blackboard site.  http://www.cmu.edu/blackboard

40. Prerequisites  Basic programming skills.  Provided by 15-111 or 15-200.  Basic discrete math and logic.  Provided by 21-127.  This course is not about Java programming, but about how to solve problems on a computer.

41. How will 15-211 be different?  Faster pace.  Lots of reading, thinking, programming.  Assumption of mature work habits.  Starting early, overcoming obstacles.  Larger homework assignments.  Typically requiring several days of work.  More open-ended problems to solve.  Sometimes no single “right” answer.

42. Textbook  Mark Allen Weiss, Data Structures & Algorithm Analysis in Java.  http://www.cs.fiu. edu/~weiss  Addison-Wesley, 2002.

43. Recitations  Starts tomorrow!  Questions are answered, new material is covered, and homeworks might be graded.  Some TAs will insist that you go to recitation in order to get a grade on your homeworks!  Switching sections is possible.  But the sections must be balanced.  Talk to the TAs – they are in charge.

44. Recitations and your TA  Recitation sessions are also the place for you to get to know your TA…  …and for your TA to get to know you  Your TA will be responsible for assigning almost all of your grades for this course!

45. Go to Recitation!!!

46. Handouts and information  Everything on Blackboard (Bb) system.  http://www.cmu.edu/blackboard  User id: your Andrew id.  Password: your Andrew password.

47. The Blackboard (Bb) system  Check Bb frequently for  Announcements.  Course policies.  You must read the cheating policy.  Contact information.  Course schedule and calendar.  Discussion bboard.

48. Requirements  Participation in lectures and recitations.  Reading assignments.  Homework assignments.  Online quizzes.  In-class and final exams.

49. Grades  1 “warmup”.  3 small homeworks.  3 large homeworks.  3 quizzes.  Midterm exam.  Final exam.  TA discretion.  39 points.  50 points each.  100 points each.  20 points each.  125 points.  275 points.  50 points. SUBJECT TO CHANGE

50. Homework assignments  Goal: Reinforce and apply what is taught in class.  Early homeworks will be small individual programming assignments.  Later homeworks will be much larger assignments, working in groups of two.

51. What kind of Assignments?  How do we write our own little search engine?  How do we compress a file?  lossless (winzip)  lossy (jpeg)  How do we design and program a game?  How to analyze the game board?  How to pick the best move?

52. Homework administration  Electronic handout (on Bb), handin, & handback.  All due on Mondays 11:59pm.  Cluster hours on some evenings, to be announced.  Up to 24 hours late for -50%.  No handins accepted after 24 hours.

53. If you need an extension  DO NOT SEND EMAIL!  All emails on this topic will be ignored.  You must see an instructor face-to- face and explain the reasons.  We will be reasonable.

54. If you need an extension DO NOT SEND EMAIL!

55. Homework assignment 0  HW0 is available now!  Download and get started today.  HW0 due on Monday at 11:59pm!  Handin of a working assignment will get the full 40 points credit.  Gets you started with the software tools used in this course.

56. Software tools  We recommend the use of standard tools.  Sun’s Java development kit (J2SE 1.4.1).  Emacs text editor.  Unix (Linux, Solaris, MacOS X, Cygwin).  Web-based handin system.

57. Cheating  There is no need to cheat in this course.  If you are in trouble, come and talk to us.  We are here to help you.  We expect you to know what is useful collaboration and what is cheating.  We will be using advanced data-mining tools to monitor the originality of all programs that are handed in

58. Original code  For programming assignments, you will be required to hand in your own original code  You may discuss your homeworks with others, but if doing so impairs your ability to write truly original code, then you are cheating  More on this as the semester goes on…

59. Consequences of cheating  If you cheat, you should come to us immediately for help.  We will try to be lenient.  If we catch you, you will either  lose at least double credit, or  fail the course.  Especially bad cases will be officially reported to the Associate Dean.

60. Homework discipline  Start early.  Most assignments will require more than 1 evening to complete.  Larger assignments will require several days of group work.  Remember:  Don’t start by sitting at a blank screen.  Humans will be trying to understand your code when grading it.

61. Summary  Go to recitation  Check out the Blackboard site  And welcome to the course!