Sunday, November 29, 2009
Sunday, November 22, 2009
Wednesday, November 18, 2009
We hope you enjoyed the graph assignment (aka Pirates). The feedback we have from you is that it was generally viewed as a worthy assignment, but that it was hard, especially for the time allotted. We want to acknowledge this, and we will grade the assignment accordingly! Additionally, several of you have asked for "how am I doing?" feedback. We are developing some information that can help you judge your overall progress, and we expect to get this to you right after Thanksgiving. We don't want you to panic unnecessarily in preparing for the final, but rather, in fact, to enjoy the experience!! -- Chris and Bill (and the TAs also)
Tuesday, November 17, 2009
Thursday, November 12, 2009
First: Your greedy bot will most likely not need/use a cartographer (if done correctly). This is okay and even expected. Second: Don't just assume that one bot(MST-TSP or Greedy) should beat the other. Actually think about which one is better for TP and why. Again, for theory question 1, note that we are telling you whatever relative behavior between MST-TSP and Greedy that you see in TP will be identical to the relative outcome in TP Redux. This statement accounts for "tricky optimizations" that you might make. --Adam
Just in case (I'd rather make sure I've said it, then have people not know), I wanted to make sure that everyone realizes that writing the GreedyCaptainBot IS a required part of the assignment. As stated in the documentation, it is tested in a private test. To make sure you are passing, you can run the MST-TSP-APPROX bot against your greedy bot (as this is exactly what we are doing!). One more note: there is nothing wrong with using a greedy cartographer in addition to your greedy captain. Hope that helps, Adam
Wednesday, November 11, 2009
Tuesday, November 10, 2009
Sorry for the delay in getting these up! There seems to be a lot of confusion in how to do MST-TSP-TOUR, so I wanted to officially make a few clarifications. (1) The graph you recieve will not necessarily be complete. This is the point of "completing the graph." To "complete the graph" your edges in the graph should be minimal path weights. This is convieniently where Dijkstra comes in. (2) To deal with a subset of vertices, you should just make a complete graph ON THOSE VERTICES, before creating an MST ON THOSE VERTICES. (3) The vertices in the list that you return represent a list of destinations. Imagine that you are at the vertex you want to start from. The list of vertices tells you where to go next. So, as a result of this, the first vertex in the list is NOT the goal, but the last is. Hope that helps! --Adam
There was a mistake in the dynprog rubric for 1 (d), consequently, stating that some rooms e_i at times t_j were inaccessible if "n+1-i > T-1-j" for the n and T given in the problem intro (or an equivalent condition) was sometimes marked off a point, even though it is accurate. If you experienced this specific error in grading, then turn back in your graded dynprog homework to your TA next recitation (November 11th) and this specific error will be corrected. If for some reason your homework is not resubmitted that recitation, you can get into contact with Trevor Burns (email@example.com) to work it out, preferably by going to his office hours (Mondays 7-8pm GHC 4122) with the homework in hand. For any other grading concerns you should pursue the normal avenues of recourse. Thanks for your patience.
Monday, November 09, 2009
Friday, November 06, 2009
I have made a few changes that shouldn't affect most people, but I'm letting you know about them anyway. (1) The compareTo code for WeightedEdge has been changed to actually take the weight into account. (2) The staff tests have been fixed to rely less on the graphs you guys make, and now use the staff graph to test your algorithms. (3) The staff code now takes self-loops into account; you will not be penalized if you do not. --Adam
Thursday, November 05, 2009
Hi, I will cancel my office this week and instead I will have my office hour for two hours (4 PM to 6 PM) in the next week. This is a onetime change due to a on-going project in our group. I'm sorry about this inconvenience.
Wednesday, November 04, 2009
Hi all, We have decided to move the due date of Traveling Pirates forward two days. A few notes about this: 1) This is NOT a reason to delay starting the assignment; in fact, the reason we are extending the time you have for the lab is that it is non-trivial. 2) This has nothing to do with FrontDesk downtime; if it does occur, we will take this into consideration (and possibly move it forward further). Do not count on FrontDesk dying though. 3) The release date of Chess will NOT be moved forward; that is, you will simultaneously have two labs out for a few days. Good luck on the exam tomorrow!!!!! --Adam
Sunday, November 01, 2009
Hey guys! Since this is a new lab, there were bound to be mistakes. A student pointed out that something didn't look quite right, and we have now fixed the error. If you re-download the stubs, you will get the change, but for those of you who don't want to, go to ( Here ) to see the exact change to the code. Thanks for the patience!