Become insanely obsessed with a problem called P=NP?
In the next 24 hours the question will be resolved....
Let it begin.
Name:
Anonymous2008-05-15 0:34
read SICP
Name:
Anonymous2008-05-15 0:35
>>2
Car Cdr Once you let the camel's heads in the tent etc
Name:
Anonymous2008-05-15 0:40
I have the entire thing conceptually done, but putting it into reality is hard and bothersome :(
Name:
Anonymous2008-05-15 0:52
I have no idea what a polynomial is and can't relate. However, you can always remind yourself that 'hard and bothersome' actually means 'I'm lazy'.
Name:
Anonymous2008-05-15 1:05
>>5
Have you ever delved into the depths of abstraction? Have you see the symmetry of an impossible maze. Have you seen that even though it is impossibly tangled at the same time it impossibly perfect? Have you battered it around into thousands of forms and torn it all apart. Have you seen the shape of the triangle of an NP complete problem? Just imagine ever exponentially growing possibilities that are perfectly symmetric about so many points yet are impossible to predict.
The problem isn't that lazy part but rather how do you translate something so beautiful into code that tames it?
Name:
Anonymous2008-05-15 1:42
Determinism is for suckers.
Name:
Anonymous2008-05-15 2:02
My friends, I do believe that we have been trolled exponentially.
The funny thing about some anonymous faggot from /prog/ solving P=NP would be that everyone could go like;
No, this may surprise you, but I proved that P=NP.
But then it would be posted on reddit, and the board would go to hell. So let's not do that.
The closest analogy I can find to the algorithm, is the maze being unwinded with string interconnecting every single movement. In order to solve it polynomially you have to basically make everything synergistic and related.
Name:
Anonymous2008-05-15 17:21
Also I am really fucking lazy and play Warcraft 3 instead of coding it in.
Name:
Anonymous2008-05-15 18:31
>>17
Many scientist have tried to prove P=NP and none of them succeeded. Some retard who has read SICP isn't going to solve it.
>>20
Many smart people have tried and failed. That indicates that smart people can't prove it.
Thus, the logical conclusion is that we should let a retard have a go at it.
>>22
I like the way you think, and would like to interview you.
Name:
Anonymous2008-05-15 19:17
>>22
This is what drives me, and thought of patenting,copywriting it, and generally weaseling every cent from it while mocking /prog/ and the internet.
EPIC OMG EPIC ST ][ OMG WI
bbcode failure.
even your automated tool failed, you failing sack of shit.
Name:
Anonymous2008-05-15 22:52
Shit, /prog/, shit. I didn't know anything about this math/computer science[1]/whatever crap. I am really fucking tired right now, but I read it anyway. Now, my brain is going to explode. Why are you doing this, WHY?
[1] i just write code, i didn't have any idea about this theory crap.
You have a set of numbers. The most complex case is when you are searching for a combination of half the set that equals the sum. So if there are 10 numbers the most complex cases are when the sum is a combination of 5 of the numbers (this can be easily seen looking at pascal's triangle.
Well when Z = the amount of numbers in the combination equally the sum. You create Z arrays each of about length Z. You can divide this into Z different sets of Z length z/2 arrays. After this step you can create Z-X * Z-Y (where X and Y sum to Z) combinations of Z length Z/4 arrays. This repeats with exponential growth following Z Z^2 Z^4 Z^8 etc. At each step a small amount of the possibilities are stricken from further combination by analyzing the midpoints of each array and throwing out all combinations higher or lower based on the result. Unfortunately because of all the independent shit happening you can only eliminate so many.
Eventually you end up with in the worst case scenarios Z arrays each of length 2 and completely indepent from eachother leaving 2^Z possibilities. From here you can eliminate possibilities pretty fast because arrays of 2 possibilities with no correspondence allow fast analysis. Unfortunately there is an exponential amount of possibilities creating at each step..
so it takes log2(n) steps obviously to reach this 2 possibility utopia of analysis.
But at each step it grows exponentially. N*n^2*n^4 etc.
So basically this results in with the knowledge that N is equievalent to 2^log2N around the same time as 2^N.
Basically it doenst solve shit and was a big waste of time.
The method I used was basically an extension of binary search. If you inserted N numbers and said a combination including only 1 of them the code would reduce this to the simple binary search case of 1 combination of N numbers.
I think some possibility exists that analyzing the difference between numbers in the sets and doing a search using these values instead of actual spots and midpoints could *maybe* (1/inf) lead to a polynomial result but basically unless you came up with a gift from the god's algorithm or something extremely novel it isn't happening soon.
The result that no one has found a polynomial time algorithm IS absolutely NOT evidence that P!=NP. The complexity of all the different possible ways of solving a question leads credence to the fact that one novel way might allow it to be solved in a polynomial way. Polynomial can still be extremely slow.
Name:
Anonymous2008-05-17 1:27
btw I didn't spend the 6 months on the above algorithms only a week or so.