hi!
i did not know that!! thanks for replying..

it is not necessary that you have to spilt the number into 2 numbers
only...
for example for '6' the nos. are (1,2,3), which give lcm as 6.
and i am not sure if it is possible to form and solve subproblems.(i
might be wrong)

but then, that was very valuable info.. i could not have guessed it
myself.
if i am over looking something obvious please let me know..

thanks!!

On Dec 3, 8:41 pm, James Fang <[EMAIL PROTECTED]> wrote:
> You needn't brute force the possible combinations.
>
> 1) if the number is odd, the lcm = ground(number/2) * (ground(number/
> 2)+1)
>
> 2) if the number is even, and number/2 is still even . the lcm =
> (number/2-1)* (number/2+1)
>
> 3) if the number is even, and number/2 is odd. then lcm = (number/
> 2-2)*(number/2+2)
>
> it can be proved as following:
> in situation 1),  let ground(number/2) = a1*a2*...*an , let
> ax=multiply of random combination of a1,a2,...,an, then it is
> deducable that (ground(number/2) +1)%ax=1
> so the lcm of ground(number/2) and (ground(number/2)+1) is
> ground(number/2) * (ground(number/2)+1).
>
> in situation 2), similar with situation 1), the only common divisor
> (number/2-1) and (number/2+1) have is 2. Since we've already know that
> number/2 is even, we can get the conclusion.
>
> similar deduction on situation 3).
>
> Regards and  Thanks,
> James Fang
>
> On Dec 2, 7:35 pm, malicious code <[EMAIL PROTECTED]> wrote:
>
> > hi!
> > suppose i have a number, say.. '7'
> > i want to find the highest possible lcm of the set of numbers that add
> > upto 7
> > that is.. in this case
> > 3+4 = 7
> > lcm(3,4) = 12
>
> > how do i go about this?
> > i could always brute force the possible combinations(1,6)(2,5)...
> > (1,2,4)....(1,1,1,1,1,1,1)
> > this is clearly very inefficient
>
> > what if i enumerate pairs of coprimes(2,3)(2,5) etc...
> > could this lead to a better solution?
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups 
"Algorithm Geeks" group.
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to [EMAIL PROTECTED]
For more options, visit this group at http://groups.google.com/group/algogeeks
-~----------~----~----~----~------~----~------~--~---

Reply via email to