Exam Post-Mortem

So how'd you all find it?

• State Space - Damn puzzles in exams, but we knew it would happen. Didn't include the boat, which is why I couldn't get it.
• CSP - Assignment questions which I was truly amazed by, so I knew it from the beginning of the year
• Resolution Refutation - Not too bad, a few statements I didn't think needed to be there, so hopefully I'm right there
• Search - Woop woop
• Alpha-Beta - 1x Beta cut, 2x Alpha cuts as far as I could tell
• CNF - Wasn't too bad
• Decision Trees - Pretty straightforward

I think I may pass, which would be great considering the work put into this and interest in it in general.

I thought it was quite difficult, i don't know how i did though...
The puzzle state space was quite a challenge, I remember seeing the problem while doing assignments.
The resolution refutation question was ok, I also found most of the statements to be useless and only used 3 or 4 of them i think.
Hopefully my attempt at the alpha-beta will get me some marks because I know I didn't get it 100% right.

Anyway, 1 more down, 2 to go still. COS3216 books and notes await

Eish, this was humbelling...
Who's joining me next hear?

Reading through the exam before-hand, and even now after the exam - I didn't think it was that difficult. It however, totally owned me. I knew the work, but for some reason I just could not figure out some of the questions.
I may just scrape though..I can hope..

@brettc - Think we've done pretty much the same subjects all the way through so far. I'm guessing you've got COS3216 and INF305 left?

I am officially taking today off though, will get onto the next horse tomorrow - think a days break is deserved after 6 exams.

Most of the questions were slight variations from tut questions.

Question 1 was straight forward. If you understood the solution to the assignment 2 question regarding the farmer/goat/cabbage/wolf, the question was easy. You didnt have to include the boat to solve the question. The search tree was long, but most of the time only one option was viable, so it wasnt too bad. Not having to include bad decisions made it much easier.

CSP: Easy easy question. And Bonus marks too! Strangely, I didnt see Q2.2 (about MRV) until checking my paper, with 5 mins left in the exam So I had to squeeze the answer in one line.

Resolution refutation: Thought it wasnt bad. Since we didnt have to decode the sentences, it made it easier. I too only used about 3-4 of the given premise. I think I got NIL on line 12.

Search: Much guess work for me. I didnt go over this. So I did what I thought was correct and excluded it from my conservative counting of marks.

Alpha-Beta: Wasnt too bad. I got two Beta cut-offs though. But I am not 100% comfortable with alpha-beta. If I am not correct, I think I will atleast get 80% of the marks.

CNF: Think I got it right, but may have made 1 minor error. If I did, I will get most of the marks.

Decisions trees: I didnt know this part of the syllabus terribly well.


With a very conservative counting of marks, I think I will just pass (either that, or only just get a supp on the higher end).

Well I struggled, but I knew I was going to, I also didnt include the boat in the first question, and had a HUGE search tree simply because by having one OR two ppl in the boat was irritating. I was hoping for a decision tree like the one in the exam tut. I used most of the statements for the Resolution Refutation except for food(apple) and food(???) dont remember the other one.

@Kyle,

Yeah we've done mostly the same subjects, I did INF305 last year though. I have COS3216 and COS361F left

I have COS361, COS321 and INF305 left.

Regarding the cannibal search tree, most of the time if you didnt travel with one person, either side could violate the M >= C condtion, and thus the option could be left off the tree.

I thought the exam was fair. If you had gone over the respective sections and knew them to a mediocre level, the question was easy. Luckily for me, I think going over the tuts and sample exam question a few times was all that was required.

I hope I am not wrong :<

For the resolution question,
I showed that peanut was a food, by using eats(bill,peanuts) with another premise that had eats(x,y) OR cant remember OR food(y)

Then there was a NOT(cant remember) to eliminate the cant remember, which left food(peanuts).

Then from the goal there was NOT(eats(john,peanuts)), and I had shown that eats(john,peanuts) somewhere, so those cancelled to prove the goal.

@Brett - Nice one, I enjoyed COS261, but decided one logic subject was enough torture and decided to go for something else instead, hehe. Goodluck for the rest (and everyone else)!

@Scag - I assumed you needed the boat from when I googled it when I got home (http://www.jimloy.com/puzz/cannibal.htm). I even made a note on the exam saying I probably just didn't see the correct sequence, because at any time that you went with two over the river (of the same type), you left a M with 2+ C's. I was more concerned with getting to the end of the paper though, so drew 2x half-paths down, and then left space to get to it after that.

I can't remember exactly what I did with the RR, but that looks about the same as what I did.

With A-B, I cut off on the left side of the tree one of the options at the bottom MIN node, and cut off 2x Tree's near the top of the right MAX node. Time will tell I guess, I'm just happy this nightmare subject (that I love) is over

@Kyle: Two thirds of COS361 is the same material as COS261 , and its only 3 chapters from the prescribed textbook: propositional logic, predicate logic, and modal logic & Agents. You should have done it.

The boat was pretty much a constant, so I dont know why it would need to be added. I solved the problem with the same {}{} technique as in the solutions to assignment 2. I encountered the problem with moving two of the same type over a few times, but managed to overcome the problem. I think I took a few moves to get two C's on the other side, then slowly replaced them with M's, finding myself in the state {C,C,C}{M,M,M}, and then the problem was easy.

From {C C C} {M M M} I thought the problem was lost, because if you sent the boat back with a Missionary, you would have {C C C M} {M M}. Ah well, I knew I was wrong, but I did use the same {} {} technique, so I'm sure there will be some sort of marks there for me. I don't mind not doing the COS361, I took some other cool subjects I wouldn't have otherwise taken, so I'm cool with that.

I think at that point the boat was on the {CCC} side, so I took two C's across, left one behind, and went back to fetch the last C.

I think the boat was on a piece of string so they could pull it from one side to the other.... that was they didnt need a driver Ha ha, this silly module!

Think I took a dive on this one... should get a supp and try in Feb 2011.

Anyway, in terms of the Missionaries and Cannibals - if you make the state space something like { C,C,C,M,M,M } or varieties like { C,C,C } then you introduce order into the system then the states explode.


How the faak do to you CNF

(p->1)->(!r -> (s^t))

I somehow ended up in DNF instead and then demorgan'd the thing