OJ Board

Ya, the problem was set as an easy dp problem. So, the constraint is set so that the obvious state dp solution can pass the time limit. If you have any question about that then well, no one is stopping you from getting it accepted by using the greedy solution. So, there shouldn't be any problem.

This very important line was missing from the problem statement (also the sample output doesn't depict this last moment change as well) - "print 3 digits after decimal point in output." Sorry for the inconvenience. We'll try to rejudge the contest submissions with a special corrector program if poss...

Ingenious, no doubt about that ...IMHO it has a good chance of being what you are probably thinking cause as soon as I saw 'PG the evil', only one ellaboration struck my mind, which I think is very much obvious.

Final will be on Tokyo.
http://icpc.baylor.edu/icpc/finals/default.htm
Btw, have a look at the schedule -
http://icpc.baylor.edu/icpc/finals/Schedule.htm
Won't there be any java challenge / parallel challenge this year? Nothing is mentioned in the schedule.

Thanks for your concern. I'm mailing the code to you. Please have a look...

Actually I was stunned to know that the judge solution was written using backtracking, too. But, in fact this is a variant of the assignment problem which has 2 solutions as far as I know. One is the classical flow solution & the other one is a backtracking solution which continously prunes the sear...

Yes, it is (source -> sum of rows) : (rows -> columns) : (sum of columns -> sink). Btw, what did you set as (rows->columns) capacity value?

I think you are right. I didn't take part in the contest cause I participated in the real-time contest held with this problemset last month. But I solved this problem during the real-time. And now, amazingly the simple Ford-Fulkerson code with that very idea is getting WA.

This problem is leaving me completely dumb. Already got about a dozen WAs today though I miserably failed to find a possible reason. It will be greatly helpful if someone who has gotten AC in this one can provide me some sample I/O.

I'm an idiot. Thanks a lot

As a reference, an ancestor query for a tree can be defined as below - ancestor(u,v) = 1, if v is an ancestor of u; -1, if u is an ancestor of v; 0, otherwise; Now, can anyone tell me that if there's a way to answer ancestor queries in O(1) time. Available pre-processing time for an n-node tree is m...

Nice job.

That means your program somehow manages to get into an infinite loop which I think not very unlikely in this input format. :roll: You can try my peculiar input strategy for this one. scanf(" %[^ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz]",buf); while(scanf(" %[ABCDEFGHIJKLMNOPQRSTUVWXYZabc...

Umm, I think I'd better show you several patterns of compare_function to use with qsort() library function. Then you yourself will be able to pick what you actually need. 1. Just beginning from the scratch, suppose you want to sort int a[10]; in ascending order. Your compare function will be like th...