| |
| |
| |
|
Page: 1 2 3 4 5 6 7 8 9 10 11
Comments:
<0> yeah,looking by this point its right.
<1> Klindel I checked in my math encyclopedia, no refereence to "linear permutations"
<1> what you've said doesn't mean anything
<2> when I say linear I just mean sequential, when people hear permutations they think any order, i.e. 1342 is a permutation of 1234 but I don't want that permutation
<1> why is {1,2}{3,4,5} sequential and {1,2,3}{4,5} not?
<2> no they are, I need an algorithm that can test all sequential permutations of K groups
<2> I cant come up with something that'll test all ssequential, only certain ones
<1> yuo aer given k groups?
<2> ya, I'm told I need to find the min cost permutation of K groups
<1> ok, you're doing math that I never studied
<2> I never studied it either, its for a dynamic programming project :-/
<1> and the problem is stated how
<2> let me link it, it'll make more sense then me explaining it
<2> http://www.cs.ucsb.edu/~teo/cs130b.w06/proj.pdf section 2.3
<1> hopefully, it makes _no_ sense now
<1> how do you like Santa Barbara?
<2> school or town?
<2> the town is nice, almost always great weather
<1> I've never been actually on the campus. I had a client there and used to fly (own airplane) from Orange County to Santa Barbara, so I've overflown it at rather low altitude many time
<1> s
<1> ok, so you've got a clustering problem in 2d, right?
<2> ya
<2> wow
<1> some neurobiology guys doing recordings in rat and monkey brains at U of Arizona
<1> we didn't use manhatten distance, but that's irrelevant
<2> he says "simple" so Im ***uming there is a trick to keep track of the # of groups and the start/end points of each group
<2> ya the cost is the simple part :)
<2> been stuck the past 8 hours trying to figure out something that can give me the best permutation
<1> how are permutations coming into ti?
<1> it?
<2> only way to describe the possible combinations
<1> so what method are you wanting to code now?
<2> well cost is easy
<0> thanks god the pastebin
<2> I think I have VAL done but it only does what its supposed to, i.e. return 3 different possible answers
<0> i overwrited my project
<0> got the original from pastebin
<1> I've always preferred the "clustering" algorithims
<2> I need to figure out what I can do to use val that'll give me kay(i,j) which will let me find the best clustring of k groups
<1> P-Brasil that's what version control systems are for also
<1> what sorts of N do you have?
<0> it is a signal that my brain is really tired
<2> I think he said less than 4000 points and the x,y coords will range between 0 and 500
<0> because i am tired but i'm insisting in make another program
<0> bye,and thanks for everything
<1> "That's what we're here for" 1986 Victor A. Wagner Jr.
<2> hehe
<1> ok, so you don't really want to permute 4000 things
<1> the range of the coords isn't all that important
<1> just as long as when you go to calculate the distance you don't overflow
<2> ya
<2> its supposed to be "dynamic programming with recursion"
<1> no sure why one would want to do recursion
<2> *shrug* its in the requirement, Val(i,j) is a recursive call
<2> my problem right now is I can't control group size, right now it always returns a single point as optimal, but then there are N groups
<1> you have an upper limit of groups
<1> I thought k
<2> ya, I don't know how to enforce it
<1> you don't let it make more than k gropus
<2> not sure how to do that
<2> http://www.noidea128.org/sourcefiles/16008.html thats all that I have written so far for it, dynam is the entry function that calls val
<1> 1 <= l <= k
<1> so you limit it on the cal to val
<1> err, 1 <= j <= k
<1> sorry
<2> ya but I don't remember the ranges at all, I can't enforce that there must be K groups
<2> any idea on that part I'll appreciate it
<1> no, vals args are limited
<2> I can change the functions, thats not a problem
<1> j <= k
<1> as a practical matter l can't really get larger than n-k either
<2> j <= k?
<1> that's what the thing you posted says
<1> 1 <= i <= n
<1> as well
<2> ah, in the project description?
<1> yes
<1> and I think that will limit the number of groups to k
<2> so in my val function I should have a double for loop then?
<2> one going from 1 -> N and another going from 1 -> k ?
<2> I still don't know how to find the limits of each group when I do this, something like {1,2,3} {4,5} and such
<3> damn it JB
<4> Ok
<4> if i use auto_ptr do i have to called delete to free memory or I don't have to do delete at all
<3> all right
<3> well
<3> I'm not hungry now
<3> but I'm still bored.
<3> I don't much care for video games.
<3> what the ****
<3> I'm finally home on a Sunday
<3> and Fox isn't showing the Simpsons or anything else I like.
<5> what's on?
<3> Bad Boys
<5> sounds like a dvd night for me!
<4> http://www.noidea128.org/sourcefiles/16009.html
<4> question regarding auto_ptr i'm very nooby new regarding this thing
<4> :P i don't even know about auto_ptr untill i join this chanel
<6> im just wall decoration here
<6> watch what everyone does
<4> i would like to be wall decoration too... but have no sufficiend skill yet ;D
<6> you probably know more than me
<6> but i read sometimes and try to learn something from it
<6> and if they talk about something they did before then i mix in the convo
<4> OK ... I'll just google somemore untill someone can help me
<7> auto_ptr = boo
<7> shared_ptr = yay!
<8> when will compilers start implementing the stuff in TR1?
<4> where to get this shared_ptr
<8> www.boost.org
<4> i will try this boost thing much more later
<6> see Noob peter can tell things very simple
<8> otherwise you can google for smart pointer
<8> and see what seems trustworthy
<4> Boost is good...
<7> NoobsGeek, to answer your question though, auto_ptr only defines an ***ignment operator for another auto_ptr
<6> like www.hax0r.tk
<6> :)
<7> so that it can take ownership of the pointer wrapped by the right hand side auto_ptr
<7> as auto_ptr transfers ownership in this manner, it is inherently unsafe to use with container cl***es
<4> which means I cant safely use it like the sample code i give?
<7> try otherObject.reset(new OtherCl***(params)); or otherObject(new OtherCl***(params)) in the init list
<7> auto_ptr docs: http://www.roguewave.com/support/docs/leif/sourcepro/html/stdlibref/auto-ptr.html
<7> NoobsGeek, by container cl***es, i mean the standard library container cl***es such as vector
<4> pardon me for not being able to tell what is container cl***es, thank now I have some idea what is container cl***es
<7> NoobsGeek, you don't have to appologize. we weren't all born knowing C++.
<4> it works
<9> as long as one admids one doesnt know one is respected
<9> right?
<7> indeed
<9> :)
<9> i know all about it
<7> there are plenty of things the ops don't know too
<9> i really dont think its wierd
<7> collectively, though, i think we know C++ in it's entirety
<9> C++ sounds kinda complex, judging from what i read here
<7> it's complexity is it's source of power, such as boost, and obviously it's weakness, as it's hard to understand until one "gets it"
<9> yeah, if it all adds up together
<9> but is it the case with C++ that its easy to do things very wrong and it still 'seems' to work?
<9> i mean that its easy to make things look like it works but to get something actualy work well is very though?
<4> Honestly, that is my problem, the part that easy to do wrong things and not realizing it
<9> the fact that you know where your weakness is is a great advantage
<4> unfortunately i will not have much time for C++ , after what I'm doing is done will have to go back to c#
<9> what more my problem is is im not good at focussing
<9> and not very good at working together much
<4> so whatever I learn in this few weeks will noe be followup with deep study
<9> might be valueble anyway
<4> <9> and not very good at working together much <--- programming in team, yeah thats great... better that solo ...
Return to
#c++
or
Go to some related
logs:
#linuxhelp
#c++
#london
tootho boot
#c++
backticks in vixie crontab
#chatzone
#AllNiteCafe
#linux
intel pro 2299bg
|
|