====================================================================== This file is a compilation of messages which were published to the ucd.class.ecs122a newsgroup (Spring 1996) ====================================================================== From news.ucdavis.edu!toadflax!blackj Thu Dec 12 10:37:27 1996 Path: news.ucdavis.edu!toadflax!blackj From: blackj@toadflax.cs.ucdavis.edu (John R. Black) Newsgroups: ucd.class.ecs122a Subject: Additional Optional Discussion Date: 8 Oct 1996 04:07:07 GMT Organization: University of California, Davis Lines: 24 Message-ID: <53ck1b$pnf@mark.ucdavis.edu> NNTP-Posting-Host: toadflax.cs.ucdavis.edu X-Newsreader: TIN [version 1.2 PL1] Several students have mentioned that the current optional discussions on Wednesdays 4-5pm does not fit their schedules. I therefore plan to convert one of my Thursday office hours into a second discussion; please vote for the time which best fits your schedule. If attendance is insufficient, this may revert to an office hour, but it will remain at the new time. Vote for some permutation of 1, 2, 3, 4, 5, in order of your preference from highest to lowest: 1. Tuesday 3:10pm 2. Tuesday 4:10pm 3. Wednesday 5:10pm 4. Thursday 1:10pm 5. Thursday 2:10pm Please vote only if you plan to attend. Send your vote to blackj@cs by Friday, 10/11. If there is sufficient interest, the new discussion will begin meeting the week of 10/14, and my Thursday office hours will be reduced from 2-4pm to 3-4pm. John Black TA 122A >From news.ucdavis.edu!usenet Thu Dec 12 10:37:27 1996 Path: news.ucdavis.edu!usenet From: Phil Rogaway Newsgroups: ucd.class.ecs122a Subject: homework clarification: problem 2-5 Date: Tue, 08 Oct 1996 14:01:04 -0700 Organization: University of California, Davis Lines: 12 Message-ID: <325AC110.41C6@cs.ucdavis.edu> NNTP-Posting-Host: rogaway.cs.ucdavis.edu Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 3.0 (X11; U; IRIX 5.3 IP22) Problem 2-5: The number of flips is a worst-case bound. This wasn't made clear in the problem's wording. Your hardworking TA has bested my solution to this problem, using a computer-assisted analysis. After much hard work (by John and his patient computer) he can make do with 2n-6 flips for n>=7. Congradulations to John on the best pancake-flipping solution in UC history! Phil Rogaway >From news.ucdavis.edu!usenet Thu Dec 12 10:37:28 1996 Path: news.ucdavis.edu!usenet From: Phil Rogaway Newsgroups: ucd.class.ecs122a Subject: possibly helpful book Date: Tue, 08 Oct 1996 15:33:04 -0700 Organization: University of California, Davis Lines: 8 Message-ID: <325AD6A0.446B@cs.ucdavis.edu> NNTP-Posting-Host: rogaway.cs.ucdavis.edu Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 3.0 (X11; U; IRIX 5.3 IP22) I've put a couple copies of Sara Baase's book "Computer Algorithms: Introduction to Design and Analysis" in Shields on 24-hour reserve. Perhaps this book (which has sometimes been used in this class) may be useful to some of you who might want to consult some text in addition to CLR. Phil Rogaway >From news.ucdavis.edu!toadflax!blackj Thu Dec 12 10:37:28 1996 Path: news.ucdavis.edu!toadflax!blackj From: blackj@toadflax.cs.ucdavis.edu (John R. Black) Newsgroups: ucd.class.ecs122a Subject: Error in PS2 solutions Date: 11 Oct 1996 23:43:03 GMT Organization: University of California, Davis Lines: 6 Message-ID: <53mm27$i6u@mark.ucdavis.edu> NNTP-Posting-Host: toadflax.cs.ucdavis.edu X-Newsreader: TIN [version 1.2 PL1] In problem 2-3 (the coin weighing problem) the final answer of log base 3 of n should have a ceiling symbol around it, not a floor. Sorry about the typo (my fault). John Black TA 122a >From news.ucdavis.edu!usenet Thu Dec 12 10:37:28 1996 Path: news.ucdavis.edu!usenet From: Phil Rogaway Newsgroups: ucd.class.ecs122a Subject: office hours Date: Mon, 14 Oct 1996 17:04:33 -0700 Organization: University of California, Davis Lines: 7 Message-ID: <3262D511.41C6@cs.ucdavis.edu> NNTP-Posting-Host: rogaway.cs.ucdavis.edu Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 3.0 (X11; U; IRIX 5.3 IP22) I have rescheduled my under-utilized office hours to: T 2-3, W 1:30 - 2:30. If you forget, it's on our Web page ... Phil Rogaway >From news.ucdavis.edu!usenet Thu Dec 12 10:37:28 1996 Path: news.ucdavis.edu!usenet From: Phil Rogaway Newsgroups: ucd.class.ecs122a Subject: ps1 results Date: Mon, 14 Oct 1996 17:32:42 -0700 Organization: University of California, Davis Lines: 6 Message-ID: <3262DBAA.167E@cs.ucdavis.edu> NNTP-Posting-Host: rogaway.cs.ucdavis.edu Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 3.0 (X11; U; IRIX 5.3 IP22) A histogram for ps1 is on our Web page. The file will grow to comprise histograms of the other homeworks, mt, etc.. Phil Rogaway >From news.ucdavis.edu!toadflax!blackj Thu Dec 12 10:37:29 1996 Path: news.ucdavis.edu!toadflax!blackj From: blackj@toadflax.cs.ucdavis.edu (John R. Black) Newsgroups: ucd.class.ecs122a Subject: No other optional discussion Date: 16 Oct 1996 03:11:06 GMT Organization: University of California, Davis Lines: 10 Message-ID: <541joa$7mn@mark.ucdavis.edu> NNTP-Posting-Host: toadflax.cs.ucdavis.edu X-Newsreader: TIN [version 1.2 PL1] Due to underwhelming demand (only 3 people voted), there will not be an additional option discussion section. If you forgot to vote and you feel strongly about it, impress upon me the urgency of your feelings (via email). Otherwise, things will stay as they are. john// ------------------------------------------------------ recursion (re ker' zhin), n. 1. see recursion >From news.ucdavis.edu!usenet Thu Dec 12 10:37:29 1996 Path: news.ucdavis.edu!usenet From: Phil Rogaway Newsgroups: ucd.class.ecs122a Subject: ps3 typos Date: Wed, 16 Oct 1996 14:16:25 -0700 Organization: University of California, Davis Lines: 6 Message-ID: <326550A8.2781@cs.ucdavis.edu> NNTP-Posting-Host: rogaway.cs.ucdavis.edu Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 3.0 (X11; U; IRIX 5.3 IP22) 3-2. Should read "T(n) <= c n" (not "T(n) <= c lg n"). 3-4. (a), (b): The "T" should be an "S". (a). Prove that the set T is NOT necessarilly unique. A corrected version of ps3 has replaced the original on the Web. >From news.ucdavis.edu!usenet Thu Dec 12 10:37:29 1996 Path: news.ucdavis.edu!usenet From: Phil Rogaway Newsgroups: ucd.class.ecs122a Subject: lecture clarification Date: Wed, 16 Oct 1996 16:48:12 -0700 Organization: University of California, Davis Lines: 42 Message-ID: <3265743C.446B@cs.ucdavis.edu> NNTP-Posting-Host: rogaway.cs.ucdavis.edu Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 3.0 (X11; U; IRIX 5.3 IP22) A certain formula was less-than-clearly justified in class; let me try again. ............................................................... Assume collision resolution by chaining and uniform hashing. What is the expected number of comparsions used to search for a random item, ITEM, assuming this item to be IN the dictinary. ............................................................... Observe: When the FIRST item was inserted into the dictionary the length of its chain (ie., the chain it landed on) was 0. When the second item was inserted into the dictionary, the *expected* length of ITS (the chain it landed on) was 1/m. When the third item was inserted into the dictionary the expected length of its chain was 2/m. ... When the n-th item was inserted into the dictionary the expected length of its chain was (n-1)/m. So: when a random one of these n items was inserted, the expected length of its chain is just the average of 1/m, ..., (n-1)/m, i.e. (1/n) [ 1/m + 2/m + ... + (n-1)/m ]. Now E[# of comparisions to serch for ITEM] = 1 + E [length of the chain where ITEM was inserted at the time that ITEM was inserted] n-1 --- = 1 + (1/n) \ (i/m) . / --- i=1 Then easy manipulation gives us that this is <= 1 + n/m - 1/2m. >From news.ucdavis.edu!usenet Thu Dec 12 10:37:29 1996 Path: news.ucdavis.edu!usenet From: Phil Rogaway Newsgroups: ucd.class.ecs122a Subject: selection recurrance Date: Thu, 17 Oct 1996 00:58:21 -0700 Organization: University of California, Davis Lines: 31 Message-ID: <3265E71D.794B@cs.ucdavis.edu> NNTP-Posting-Host: rogaway.cs.ucdavis.edu Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 3.0 (X11; U; IRIX 5.3 IP22) Here's the recurrance from lecture for our deterministic selection algorithm, with a brief explanation of where each term comes from. I'm ignoring "round off" issues, just as I did in class, and as you may on your homework problem relating to this. T(n) = 6n/5 + T(n/5) + 2n/5 + T(7n/10) To compute the To recursively To separate the elements To recursively medians of each find the median which are less than the do the selection 5-some, using the of the medians median-of-medians from the on the appropriate (claimed) algorithm elements which are greater side of the partitioned that finds the median than the median-of-medians: elements. (This is of 5 elements using 6 3n/10 elements are *known* a pessimistic count for comparisons less than the median-of-medians; the size of the set that 3n/10 elements are *known* greater remains after removing than the median-of-medians; and the one side of the remaining 2n/5 elements need to be partitioned elemnts.) compared to the median-of-medians. >From news.ucdavis.edu!toadflax!blackj Thu Dec 12 10:37:29 1996 Path: news.ucdavis.edu!toadflax!blackj From: blackj@toadflax.cs.ucdavis.edu (John R. Black) Newsgroups: ucd.class.ecs122a Subject: Grade Reporting program available for ECS 122A Date: 18 Oct 1996 02:46:37 GMT Organization: University of California, Davis Lines: 38 Message-ID: <546r2d$otu@mark.ucdavis.edu> NNTP-Posting-Host: toadflax.cs.ucdavis.edu X-Newsreader: TIN [version 1.2 PL1] You can check your scores to-date at any time by sending mail to the account "cs122at" with a subject of "scores". There is a program on the other end which will reply within seconds and send you mail showing all your scores to-date including problem sets, midterm, and final. (Right now it's just PS1 of course.) A score report looks like this: ----------------------------------------- Report for: Black, John Desc Pts/Poss Wt ------ -------- ----- PS1 = 23/40 ( 3%) Weighted average is 57.50 percent EC Pts: * (Scores last updated: 1819 Oct 16) ---------------------------------------- Each line gives a description followed by the number of points you received out of the number possible. The weight (given in parentheses) indicates how heavily this item is counted toward your overall average. For the problem sets, there will be roughly 9 and as a group they represent 30% of your grade, so each one is worth about 3.33%. The "EC Pts" is extra credit points which are given out for special problems (like PS1 problem 2(c)) or for especially nice answers to other problems. EC pts are indicated by asterisks (1 per point earned). Please check your scores from time to time to make sure we are accurately recording them. If there is a descrepancy, bring your graded paper to my attention and I will fix it. John Black >From news.ucdavis.edu!usenet Thu Dec 12 10:37:30 1996 Path: news.ucdavis.edu!usenet From: Phil Rogaway Newsgroups: ucd.class.ecs122a Subject: deterministic selection Date: Thu, 17 Oct 1996 19:50:13 -0700 Organization: University of California, Davis Lines: 51 Message-ID: <3266F065.15FB@cs.ucdavis.edu> NNTP-Posting-Host: rogaway.cs.ucdavis.edu Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 3.0 (X11; U; IRIX 5.3 IP22) > From heinig@cs.ucdavis.edu Thu Oct 17 18:29:49 1996 > Date: Thu, 17 Oct 1996 18:23:55 -0700 (PDT) > From: Lael Heinig > To: rogaway@cs.ucdavis.edu > Subject: Hw 3 > Mime-Version: 1.0 > Content-Type: TEXT/PLAIN; charset=US-ASCII > > Professor Rogaway, > > A number of us have been confused on the amount of time for the > partition step of the Deterministic Selection problem. John said that he > thinks it should take n/5 for 5 elements. You have said that it takes > 2n/5. If this is true would 3 elements take n/3; 7 elements take 3n/7 > and so on. Or would 3 elements take 2n/3; 7 elements take 2n/7 and so > on. This is where the confusion lies. Therefore we don't know what the > correct answers are to problem #1. If you could shed some light on this > problem/step, we would appreciate it. > > > Thank you, > Lael Heinig I had in class: T(n) = 6n/5 + T(n/5) + 2n/5 + T(7n/10). The 2n/5 term was for the comparisons needed to partition the n elements on the median-of-the-medians (step 4 on p. 190 of the text). There are 3n/10 elements which are KNOWN smaller than the median-of-the-medians; there are 3n/10 elements which are KNOWN larger than the median-of-the-medians; and that leaves n - (3n/10 + 3n/10) = 2n/5 elements whose relation to the median-of-medians is unknown. These we should compare with the median-of-medians to accomplish the partitioning. I believe the n/5 term John was thinking about was to compare the median of each column with the median-of-the medians -- ie to rearrange our columns with the columns whose medians are less than the median-of-medians on the left, and the columns whose medians are greater than the median-of-medians on the right. But actually we don't need to invest any new comparisons to accomplish this task: in identifying the median-of-the-medians we will have necessarily determined which columns had medians which are less than the median-of-medians and which columns had medians greater than the median-of-medians. Else we couldn't have correctly determined the median-of-medians. -Phil Rogaway >From news.ucdavis.edu!toadflax!blackj Thu Dec 12 10:37:30 1996 Path: news.ucdavis.edu!toadflax!blackj From: blackj@toadflax.cs.ucdavis.edu (John R. Black) Newsgroups: ucd.class.ecs122a Subject: New Office Hours Schedule Date: 20 Oct 1996 20:40:05 GMT Organization: University of California, Davis Lines: 14 Message-ID: <54e2n5$c3b@mark.ucdavis.edu> NNTP-Posting-Host: toadflax.cs.ucdavis.edu X-Newsreader: TIN [version 1.2 PL1] A large number of students have complained that my only office hours are on Thursdays, the day before the problems sets are due. To better accommodate those who begin earlier on the homework and need help earlier in the week, I am moving one of my Thursday hours to Tuesday. The remaining Thursday hour is left skewed so that people with 1:40-3pm classes or 3:10-4:30pm classes will still have access. New Office Hours for TA: Tues 4:10 - 5pm Thurs 2:30 - 3:30pm I hope this works out better for most of you. John Black >From news.ucdavis.edu!toadflax!blackj Thu Dec 12 10:37:30 1996 Path: news.ucdavis.edu!toadflax!blackj From: blackj@toadflax.cs.ucdavis.edu (John R. Black) Newsgroups: ucd.class.ecs122a Subject: PS4 Date: 22 Oct 1996 01:03:44 GMT Organization: University of California, Davis Lines: 14 Message-ID: <54h6hg$ll5@mark.ucdavis.edu> NNTP-Posting-Host: toadflax.cs.ucdavis.edu X-Newsreader: TIN [version 1.2 PL1] In attempt to better serve you, I've done all the homework that's due Friday and I could find no typos, so have at it! :) ADDITIONAL REQUIREMENT FOR 4-1: Dr. Rogaway didn't give a maximum size for m (the hash table size); please make sure your m is less than 200. This shouldn't be too hard as there are only 48 words so we're letting you use 4 times the space in the hash table. I was able to find LOTS of functions that worked. Your hash function should be "short and simple". This means exactly what you'd expect (it's just hard to describe what you expect). John >From news.ucdavis.edu!usenet Thu Dec 12 10:37:30 1996 Path: news.ucdavis.edu!usenet From: Phil Rogaway Newsgroups: ucd.class.ecs122a Subject: Problem 4-4 postponed Date: Tue, 22 Oct 1996 19:02:00 -0700 Organization: University of California, Davis Lines: 10 Message-ID: <326D7C98.41C6@cs.ucdavis.edu> NNTP-Posting-Host: rogaway.cs.ucdavis.edu Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 3.0 (X11; U; IRIX 5.3 IP22) I didn't get quite as far as I expected on Monday, and so you may still be unclear on amortized analysis. Thus I'll postpone Problem 4-4; please turn it in with next week's homework. Hope that helps! Phil Rogaway ps: Happy hash hunting! >From news.ucdavis.edu!toadflax!blackj Thu Dec 12 10:37:31 1996 Path: news.ucdavis.edu!toadflax!blackj From: blackj@toadflax.cs.ucdavis.edu (John R. Black) Newsgroups: ucd.class.ecs122a Subject: Splay Tree Demo Date: 29 Oct 1996 02:48:14 GMT Organization: University of California, Davis Lines: 10 Message-ID: <553r9e$eoi@mark.ucdavis.edu> NNTP-Posting-Host: toadflax.cs.ucdavis.edu X-Newsreader: TIN [version 1.2 PL1] As we learn the wonderous ways of the splay tree you may want to supplement your intuition by playing with these things a little. The co-inventor of splay trees, Daniel Sleator, provides a demo on his homepage. Next time you're waiting for that long compilation to finish, take a peek at http://gs213.sp.cs.cmu.edu/cgi-bin/splay John Black (Serving you since 1996) >From news.ucdavis.edu!usenet Thu Dec 12 10:37:31 1996 Path: news.ucdavis.edu!usenet From: Phil Rogaway Newsgroups: ucd.class.ecs122a Subject: midterm - review session Date: Wed, 30 Oct 1996 13:48:09 -0800 Organization: University of California, Davis Lines: 2 Message-ID: <3277CD19.41C6@cs.ucdavis.edu> NNTP-Posting-Host: rogaway.cs.ucdavis.edu Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 3.0 (X11; U; IRIX 5.3 IP22) Review session - Wed, Nov 6, 6-8 pm, 1120 Bainer Midterm - Fri, Nov 8, in class >From news.ucdavis.edu!usenet Thu Dec 12 10:37:31 1996 Path: news.ucdavis.edu!usenet From: Phil Rogaway Newsgroups: ucd.class.ecs122a Subject: "sharing" homework solutions... Date: Wed, 30 Oct 1996 14:15:38 -0800 Organization: University of California, Davis Lines: 20 Message-ID: <3277D38A.167E@cs.ucdavis.edu> NNTP-Posting-Host: rogaway.cs.ucdavis.edu Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 3.0 (X11; U; IRIX 5.3 IP22) It's been brought to my attention that several students have turned in homeworks that are substantially identical. I would like to remind the entire class of the standard on academic misconduct wrt the homeworks: 1. You must write up problems entirely on your own. 2. You may *discuss* problems with your peers. If you do this, you must acknowledge them. "Discuss" means sharing a general idea or approach -- it wouldn't entail doing a write-up together or working out all details of a problem. On 2: I advise *against* working with anyone else on homework problems. The process of individually struggling to solve a problem is an essential part of the learning experience for this class. The midterm and final will in part be testing how well you've been acquiring problem-solving skills. Phil Rogaway >From news.ucdavis.edu!toadflax!blackj Thu Dec 12 10:37:32 1996 Path: news.ucdavis.edu!toadflax!blackj From: blackj@toadflax.cs.ucdavis.edu (John R. Black) Newsgroups: ucd.class.ecs122a Subject: PS4 hash functions Date: 5 Nov 1996 01:50:06 GMT Organization: University of California, Davis Lines: 17 Message-ID: <55m6ge$91p@mark.ucdavis.edu> NNTP-Posting-Host: toadflax.cs.ucdavis.edu X-Newsreader: TIN [version 1.2 PL1] I am truly sorry it's taken so long to grade and return the last two problem sets. Hopefully we'll do better in the last half of the quarter. Most of you found good hash functions for PS4; however a few people submitted hash functions that the grader could not get to work (in a few cases I confirmed this). If you have a star in the upper left corner of your graded homework, this means that your hash function didn't work when tested. You were tentatively given 0 points for it. But I assume there was simply some unstated assumption you were making, or some oversight we are making that is causing the confusion; so please clarify this in the next week or so by any of the following means: (1) Show me after class what you meant (2) Come to office hours and explain (3) Send me email (blackj@cs) john// >From news.ucdavis.edu!usenet Thu Dec 12 10:37:32 1996 Path: news.ucdavis.edu!usenet From: Phil Rogaway Newsgroups: ucd.class.ecs122a Subject: ps5 solns; ps6 - part 1 Date: Tue, 05 Nov 1996 23:43:12 -0800 Organization: University of California, Davis Lines: 3 Message-ID: <32804190.41C6@cs.ucdavis.edu> NNTP-Posting-Host: rogaway.cs.ucdavis.edu Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 3.0 (X11; U; IRIX 5.3 IP22) The subject material can be found on our course Web page.... -Phil Rogaway >From news.ucdavis.edu!usenet Thu Dec 12 10:37:32 1996 Path: news.ucdavis.edu!usenet From: Phil Rogaway Newsgroups: ucd.class.ecs122a Subject: revised ps5 solutions Date: Wed, 06 Nov 1996 21:03:49 -0800 Organization: University of California, Davis Lines: 3 Message-ID: <32816DB5.167E@cs.ucdavis.edu> NNTP-Posting-Host: rogaway.cs.ucdavis.edu Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 3.0 (X11; U; IRIX 5.3 IP22) A minor correction was made to 4-4(b) and a (simpler but less efficient) solution was added to 5-3(a). Grab the new-and-improved solutions off the Web! >From news.ucdavis.edu!usenet Thu Dec 12 10:37:34 1996 Path: news.ucdavis.edu!usenet From: Phil Rogaway Newsgroups: ucd.class.ecs122a Subject: midterm problems 2, 3.1 re-submission Date: Mon, 11 Nov 1996 18:00:34 -0800 Organization: University of California, Davis Lines: 38 Message-ID: <3287DA42.2781@cs.ucdavis.edu> NNTP-Posting-Host: rogaway.cs.ucdavis.edu Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 3.0 (X11; U; IRIX 5.3 IP22) In case anyone missed class today: Due in class on WEDNESDAY: a solution to problems 2 and 3.1 from the MT. For 2.2, you need only calculate the C_{i,j} values -- you needn't compute the routing. Guidelines: -No Collaboration! -Maximum of 1 page for the two problems together (ie., half a page per problem). Grading will probably be binary: I'll decide if you got the problem or you didn't. I do not guarantee to grade both problems (I might grade only a random one). I will do the grading myself. Doing the problems correctly now will not retroactively change your midterm grade. But I will record that you were able to do the problems in the absence of time pressure, and this may have a positive influence when your overall grade is decided. I hope you enjoy getting the chance to put down what you would have liked to have written down on the midterm. This isn't some sort of "punishment" -- I just want you to solve these problems and feel good about doing that. Phil Rogaway >From news.ucdavis.edu!usenet Thu Dec 12 10:37:34 1996 Path: news.ucdavis.edu!usenet From: Phil Rogaway Newsgroups: ucd.class.ecs122a Subject: ps6 postponed until monday Date: Mon, 11 Nov 1996 18:03:20 -0800 Organization: University of California, Davis Lines: 5 Message-ID: <3287DAE8.446B@cs.ucdavis.edu> NNTP-Posting-Host: rogaway.cs.ucdavis.edu Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 3.0 (X11; U; IRIX 5.3 IP22) Because of the unexpected re-submission of the midterm questions, I will postpone the due date of PS 6 to next Monday. Hope that helps, Phil Rogaway >From news.ucdavis.edu!toadflax!blackj Thu Dec 12 10:37:34 1996 Path: news.ucdavis.edu!toadflax!blackj From: blackj@toadflax.cs.ucdavis.edu (John R. Black) Newsgroups: ucd.class.ecs122a Subject: Dynamic Programming problem Date: 14 Nov 1996 23:59:09 GMT Organization: University of California, Davis Lines: 19 Message-ID: <56gbod$lnn@mark.ucdavis.edu> NNTP-Posting-Host: toadflax.cs.ucdavis.edu X-Newsreader: TIN [version 1.2 PL1] Several people have requested a new DP problem which they could work out on their own to try and firm-up their skills. So I propose you do the following problem, then come to discussion or office hrs next week and we'll compare notes. On the first day of class, Dr. Rogaway gave the following problem: given only 4 coins with values 1, 3, 8, and 17, make change for n cents. Part I: Will a greedy alg work here? If yes, prove it; if no, show a counterexample. Part II: What is the minimizing expression for this problem (you can check your notes after you try to work it out on your own). Part III: Code up a program that solves this problem. If you want to check it out, write it in your favorite language and try it for various values of n. john// >From news.ucdavis.edu!matloff Thu Dec 12 10:37:34 1996 Path: news.ucdavis.edu!matloff From: matloff@heather.cs.ucdavis.edu (Norm Matloff) Newsgroups: ucd.class.ecs40,ucd.class.ecs110,ucd.class.ecs122a,ucd.class.ecs140a,ucd.class.ecs150,ucd.class.ecs154a Subject: survey Date: 16 Nov 1996 04:46:20 GMT Organization: University of California, Davis Lines: 20 Message-ID: Reply-To: matloff@cs.ucdavis.edu NNTP-Posting-Host: heather.cs.ucdavis.edu X-Newsreader: slrn (0.8.8.2 UNIX) Xref: news.ucdavis.edu ucd.class.ecs40:684 ucd.class.ecs110:457 ucd.class.ecs122a:187 ucd.class.ecs140a:410 ucd.class.ecs150:478 ucd.class.ecs154a:280 If you are in any of the courses listed below THIS QUARTER, please send me e-mail saying the average number of hours you spend outside of class per week. Again, this is an average, so please try to account for both slack periods and busy periods. Please E-MAIL me your response (matloff@cs). The data (overall averages, without names) will be used by our Undergraduate Affairs Committee in discussing possible curricular changes. Thanks very much! Here are the courses of interest: ECS 40 ECS 110 ECS 122A ECS 140A ECS 150 ECS 154A >From news.ucdavis.edu!toadflax!blackj Thu Dec 12 10:37:34 1996 Path: news.ucdavis.edu!toadflax!blackj From: blackj@toadflax.cs.ucdavis.edu (John R. Black) Newsgroups: ucd.class.ecs122a Subject: No TA off hrs this Thursday Date: 18 Nov 1996 20:11:17 GMT Organization: University of California, Davis Lines: 12 Message-ID: <56qft5$kau@mark.ucdavis.edu> NNTP-Posting-Host: toadflax.cs.ucdavis.edu X-Newsreader: TIN [version 1.2 PL1] I have my PhD preliminary exams for four afternoons starting Thursday, so I won't make office hrs this Thursday. Also, the following Tuesday (11/26) my exam ends at 4:30, so I will move my office hrs from 4-5 to 5-6 (expect me to be slightly incoherent (yes, more than usual)). Also, I won't be as available for email questions during this period. If you need help on the homework, remember our lonely professor holding his own office hours! :) john// >From news.ucdavis.edu!toadflax!blackj Thu Dec 12 10:37:34 1996 Path: news.ucdavis.edu!toadflax!blackj From: blackj@toadflax.cs.ucdavis.edu (John R. Black) Newsgroups: ucd.class.ecs122a Subject: PS7 due date postponed Date: 21 Nov 1996 02:49:02 GMT Organization: University of California, Davis Lines: 3 Message-ID: <570fuu$9ag@mark.ucdavis.edu> NNTP-Posting-Host: toadflax.cs.ucdavis.edu X-Newsreader: TIN [version 1.2 PL1] PS7 is no longer due this Friday, but is instead due Monday, 11/25. john// >From news.ucdavis.edu!toadflax!blackj Thu Dec 12 10:37:35 1996 Path: news.ucdavis.edu!toadflax!blackj From: blackj@toadflax.cs.ucdavis.edu (John R. Black) Newsgroups: ucd.class.ecs122a Subject: Reminder: do the DP example Date: 25 Nov 1996 19:49:57 GMT Organization: University of California, Davis Lines: 8 Message-ID: <57ct95$rh1@mark.ucdavis.edu> NNTP-Posting-Host: toadflax.cs.ucdavis.edu X-Newsreader: TIN [version 1.2 PL1] If you want to practice your dynamic programming skills, remember to do the practice problem for this Wednesday's discussion (it's the coin-changing problem with denominations 1, 3, 8, and 17). I'll be posting my solution on the newsgroup as well, for those who cannot attend. john// >From news.ucdavis.edu!usenet Thu Dec 12 10:37:35 1996 Path: news.ucdavis.edu!usenet From: Phil Rogaway Newsgroups: ucd.class.ecs122a Subject: today's optional discussion Date: Wed, 27 Nov 1996 10:56:48 -0800 Organization: University of California, Davis Lines: 3 Message-ID: <329C8EF0.41C6@cs.ucdavis.edu> NNTP-Posting-Host: rogaway.cs.ucdavis.edu Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 3.0 (X11; U; IRIX 5.3 IP22) John will hold his usual discussion today, after class, where he will discuss money, marriage, and other topics of ecs122a. >From news.ucdavis.edu!toadflax!blackj Thu Dec 12 10:37:35 1996 Path: news.ucdavis.edu!toadflax!blackj From: blackj@toadflax.cs.ucdavis.edu (John R. Black) Newsgroups: ucd.class.ecs122a Subject: Using DP to solve the coin changing problem Date: 28 Nov 1996 02:08:45 GMT Organization: University of California, Davis Lines: 88 Message-ID: <57is7d$jr7@mark.ucdavis.edu> NNTP-Posting-Host: toadflax.cs.ucdavis.edu X-Newsreader: TIN [version 1.2 PL1] As mentioned today in lecture, if you're given coin denominations 1, 3, 8, and 17, you need DP to solve the problem of finding the fewest # of coins which sum to n. (Greedy doesn't work: try n=24) The minimizing expression (which Dr. Rogaway gave on the 1st day of class) was C(n) = 1 + min{C(n-1) + C(n-3) + C(n-8) + C(n-17)} Turning this into a program was not too hard. Try compiling and running and toying with this if you like, but above all else, be sure you understand the derivation: #include #define MAXVAL 1000 void c(int amt); main(int argc, char **argv) { int amt = atoi(argv[1]); int i; if (amt < MAXVAL) c(amt); else printf("amt %d is too large; limit is %d\n", amt, MAXVAL); } void c(int amt) { #define PENNY 1 #define THRENNY 3 #define EIGHT 8 #define SEVENTEEN 17 int A[MAXVAL]; int coinlink[MAXVAL]; int bsf, i, use; int pennies, thrennies, eights, seventeens; /* bootstrap process at 0 */ A[0] = 0; for (i=1; i <= amt; i++) { bsf = 1 + A[i-1], use = PENNY; if ((i >= THRENNY) && (1 + A[i-THRENNY] < bsf)) bsf = 1 + A[i-THRENNY], use = THRENNY; if ((i >= EIGHT) && (1 + A[i-EIGHT] < bsf)) bsf = 1 + A[i-EIGHT], use = EIGHT; if ((i >= SEVENTEEN) && (1 + A[i-SEVENTEEN] < bsf)) bsf = 1 + A[i-SEVENTEEN], use = SEVENTEEN; A[i] = bsf; coinlink[i] = use; } printf("The best way to make change for %d cents is to use %d coins:\n", amt, A[amt]); while (amt) { switch (coinlink[amt]) { case PENNY: pennies++; amt -= PENNY; break; case THRENNY: thrennies++; amt -= THRENNY; break; case EIGHT: eights++; amt -= EIGHT; break; case SEVENTEEN: seventeens++; amt -= SEVENTEEN; break; } } printf("%d Pennies\n", pennies); printf("%d Thrennies\n", thrennies); printf("%d 8 cent coins\n", eights); printf("%d 17 cent coins\n", seventeens); } Here is a sample run: % cc change.c -o change % change 49 The best way to make change for 49 cents is to use 5 coins: 0 Pennies 0 Thrennies 4 8 cent coins 1 17 cent coins % >From news.ucdavis.edu!toadflax!blackj Thu Dec 12 10:37:35 1996 Path: news.ucdavis.edu!toadflax!blackj From: blackj@toadflax.cs.ucdavis.edu (John R. Black) Newsgroups: ucd.class.ecs122a Subject: Weights changed for Problem Sets Date: 2 Dec 1996 02:16:02 GMT Organization: University of California, Davis Lines: 11 Message-ID: <57te52$6mm@mark.ucdavis.edu> NNTP-Posting-Host: toadflax.cs.ucdavis.edu X-Newsreader: TIN [version 1.2 PL1] For those who check their marks via the "cs122at" mail interface: The weights of each problem set used to be 3.33% each. Today I changed them to reflect the number of problems in each set; this means problem sets with 3 problems (PS4, PS7, PS8) are worth 2.9% each, those with 4 problems (PS1, PS3, PS5) are worth 3.87% each, and those with 5 problems (PS2, PS6) are worth 4.84% each. The total should add to up 30%. john// >From news.ucdavis.edu!toadflax!blackj Thu Dec 12 10:37:35 1996 Path: news.ucdavis.edu!toadflax!blackj From: blackj@toadflax.cs.ucdavis.edu (John R. Black) Newsgroups: ucd.class.ecs122a Subject: Hints for #2a Date: 4 Dec 1996 07:26:22 GMT Organization: University of California, Davis Lines: 37 Message-ID: <58392u$1h3@mark.ucdavis.edu> NNTP-Posting-Host: toadflax.cs.ucdavis.edu X-Newsreader: TIN [version 1.2 PL1] I said the following in office hrs today, so I wanted to post here for those who are stuck on #2a and didn't come to office hours. There *is* a configuration for n=3 which does not terminate. Searching haphazardly will probably not discover it; you must search systematically. Here is how I found it: Start with the marriage Aa Bb Cc and the preference lists: A B C a b c Then decide which pairs are unstable; they're all equivalent to start with, so just choose A and b (i.e. A wants b more than a, and b wants A more than B). So the new preference lists look like this: A _b_a_ B C a b _A_B_ c And we marry Ab leaving Ab Ba Cc. I left underscores above to indicate the positions which c and C could take later on; we haven't yet determined where they go and we want to be as flexible as possible because we need to keep the marriages unstable (if the marriages become stable, the algorithm terminates!). Now enumerate the possibilities from here. It takes a while, and you have to think carefully, but you will eventually find a set which cycles. Good luck... John >From news.ucdavis.edu!usenet Thu Dec 12 10:37:36 1996 Path: news.ucdavis.edu!usenet From: Phil Rogaway Newsgroups: ucd.class.ecs122a Subject: review session Date: Thu, 05 Dec 1996 13:31:12 -0800 Organization: University of California, Davis Lines: 5 Message-ID: <32A73F20.167E@cs.ucdavis.edu> NNTP-Posting-Host: rogaway.cs.ucdavis.edu Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 3.0 (X11; U; IRIX 5.3 IP22) The 122a review session will be held Wednesday, Dec 11, from 7-9pm, in 1130 Bainer. See ya there! Phil Rogaway >From news.ucdavis.edu!usenet Thu Dec 12 10:37:36 1996 Path: news.ucdavis.edu!usenet From: Phil Rogaway Newsgroups: ucd.class.ecs122a Subject: non-standard office hours Date: Sun, 08 Dec 1996 14:31:54 -0800 Organization: University of California, Davis Lines: 9 Message-ID: <32AB41DA.41C6@cs.ucdavis.edu> NNTP-Posting-Host: rogaway.cs.ucdavis.edu Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 3.0 (X11; U; IRIX 5.3 IP22) I suspect that non-standard office hours might better suit your needs this week, and so the following replaces my usual hours: W 7-9 pm - review session R 6-7:30 pm - office hours F 12:30-1:30 pm - office hours -Phil Rogaway Subject: Favorite Algorithms Date: 17 Dec 1996 01:19:28 GMT From: blackj@toadflax.cs.ucdavis.edu (John R. Black) Organization: University of California, Davis Newsgroups: ucd.class.ecs122a Here are the stats for which algorithms were the favorites during the final exam. Dijkstra's, KMP, and Kruskals all tied with 3 votes each (also I threw various flavors of "greedy" together to get another category with 3 votes). The student who specified "stooge sort" lost points for not indicating which stooge he intended. My favorite? Well, personally I want thank Dr. Rogaway for his lecture on topological sorting: since then I've rarely put my socks on after my shoes. Thanks to everyone for making our job enjoyable and rewarding; best of luck in your careers... John Black ----------------------------------------- All Pairs Shortest Paths 1 Depth First Search 1 Dijkstra's 3 Forward Radix 1 Greedy Algs 3 Heapsort 1 Huffman 2 KMP 3 Kruskal's 3 Mergesort 1 NP-Completeness Reductions 1 Program Checking Algs 1 Quicksort 1 Radiosity (?) 1 String Matching Automaton 1 Stooge Sort 1 Stable Marriage Alg 2 Z-buffer (graphics alg) 1