\[\text{ The function } f(n) \text{ is defined for all positive integers } \\ n \text{ and takes on non-negative integral values } \\ \text{ Also, for all } m, n \\ f(m+n)-f(m)-f(n)=0 \text{ or } 1 \\ f(2)=0, f(3)>0, \text{ and } f(9999)=3333. \\ \text{ Determine } f(1982).\]
https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=1&cad=rja&uact=8&ved=0ahUKEwip3YzIuqLJAhWKlh4KHUR3AcAQFggdMAA&url=http%3A%2F%2Fdb.math.ust.hk%2Fnotes_download%2Felementary%2Falgebra%2Fae_A9.pdf&usg=AFQjCNHv-dnZdBdd7dU00Pozi526wMFH6g got this question from here
under harder problems number 8 last page
@AlexandervonHumboldt2 can you please tell me what integral value means?
if you know
sorry @freckles , never heard of such term.
its k
sounds like it is just an integer
I guess I will assume it is an integer until I run into problems :p
hmmm... well I know since f(2)=0 we have \[f(2-n)+f(n)=0 \text{ or } -1 \]
and if I used f(2)=0 again... like replace n with 2 I get \[f(2-2)+f(2)=0 \text{ or } -1 \\ f(0)+f(2)=0 \text{ or } -1 \\ \text{ and this means } f(0)+0=0 \text{ or } -1 \text{ so we have} f(0)=0 \text{ or } -1 \]
shuldn't it be 0 or +1?
well I had multiplied both sides by -1 earlier
I didn't show that but I could show you what I did
oh okay
\[f(m+n)-f(m)-f(n)=0 \text{ or } 1 \\ \text{ replace } m+n=2 \text{ and so } m=2-n \\ f(2)-f(2-n)-f(n)=0 \text{ or } 1 \\ -f(2-n)-f(n)=0 \text{ or } 1 \] and then I multiplied both sides by -1
getting that one equation I had above
f(1) will be 0
I got 0 or -1/2 for f(1)
\[f(2-n)+f(n) =0 \text{or } -1 \\ \text{ replace } n \text{ with } 1 \\ \\ f(1)+f(1)=0 \text{ or } -1 \\ f(1)=0 \text{ or } \frac{-1}{2} \\ \]
but it says that the function is defined for all positive integers n and takes NON NEGATIVE values so f(1) can't be -1/2
oh I thought that meant we couldn't plug in negative integers
I didn't know it meant we couldn't have negative outputs
so if that is the case then f(0) is also 0
though I guess we couldn't use that really because the input ended up being 0 instead of positive integer
that was above when I replaced n with 2
so f(3)=1
oh my
I wonder if this is fibnooci sequence
we know that the function will give positive value we can prove that n can't be negative \[f(3-1)-f(3)-f(-1)=0or1\]\[0-ve-f(-1)=0or1\]to make LHS 1 or 0 -f(-1) has to be positive so f(-1) has to be negative and this can't happen so we can't have negative n hmm is this special case or general
definitely not Fibonacci sequence I don't think the 9999th term is 3333
and I got f(3)=1 by the following: \[f(2+1)-f(2)-f(1)=0 \text{ or } 1 \\ f(3)-0-0=0 \text{ or } 1 \\ f(3)=1 \text{ since } f(3)>0\]
yes
\[f(2+2)-f(2)-f(2)=0 \text{ or } 1 \\ f(4)=0 \text{ or } 1 \] we can't determine f(4) yet
actually I think I just determine it was 1
watch this... \[f(3+1)-f(3)-f(1)=0 \text{ or } 1 \\ f(4)-1-0=0 \text{ or } 1 \\ f(4)=1 \text{ or } 2 \]
yes :)
\[f(4)=(1 \text{ or } 2 ) \text{ and } (1 \text{ or } 0)=1\]
that f(9999)=3333 seems interesting...
tricky and confusing..nice question
yeah @Kainui kinda brought up functional equations and I said I was going to do my own research thought it would be nice to post this question since it did look tricky and fun
maybe we can try to use that cauchy equation thing on page 8
but we have a problem for the or 1 part
\[\text{ if } f(m+n)-f(m)-f(n)=0 \text{ then } f(m+n)=f(m)+f(n) \\ \text{ then } f(x)=cx \text{ where } c \text{ is a constant }\]
\[\text{ if } f(m+n)-f(m)-f(n)=1 \text{ then } f(m+n)=f(m)+f(n)+1 \\ \text{ and if } f(x)=cx+b \\ \text{ then we have } \\ c(m+n)+b=cm+b+cn+b+1 \\ c(m+n)+b=c(m+n)+2b+1 \\ b=2b+1 \\ -b=1 \\ b=-1 \] so I guess f(x)=cx-1 ...if I did that right
\[f(2)=0 \\ \text{ so } f(x)=cx \text{ or } f(x)=cx-1 \\ \text{ if } f(x)=cx \text{ then } f(2)=2c=0 \implies c=0 \text{ so } f(x)=0 \\ \text{ or we could have } \\ f(x)=cx-1 \text{ then } f(2)=2c-1=0 \implies c=\frac{1}{2} \\ \text{ and so } f(x)=\frac{1}{2}x-1\]
now f(3)>0 \[\text{ but we got } f(3)=1 \text{ and neither one fits }\]
but maybe for some values we have f(x)=0 while for other values we have f(x)=cx-1 but I solved f(3)=c(3)-1 for c and got c=2/3 and this does not work for f(9999)=3333
maybe f(x)=cx-1 isn't the most general form for f(m+n)=f(m)+f(n)+1 that could be another possibility
f(5) will be 2
shuld we find values like this or do some other approach
I was trying another approach
any approach you can think of
how did you get f(5)=2?
\[f(4+1)-f(4)-f(1)=0or1\]\[f(5)-1=0or1\] \[f(2+3)-f(2)-f(3)=0or1\]\[f(5)-2=0or1\] If the 1st eq has value 0 then f(5) will bw 1 and if this happens the 2nd equation will become -1 so f(5)=1 is rejected now we have only 1 option left in 1st eq so f(5)-1=1 so f(5)=2
but f(2)=0 and f(3)=1 so -f(2)-f(3)=-0-1=-1 not -2
): sry i made a mistake out there
don't be sad
I did find a answer. I don't know how much I like it.
http://www.cs.cornell.edu/~asdas/imo/imo/isoln/isoln821.html https://books.google.com/books?id=okx0d9jdM8oC&pg=PA144&lpg=PA144&dq=f(m%2Bn)-f(m)-f(n)%3D0+or+1+f(2)%3D0+f(9999)%3D3333&source=bl&ots=ZzeuD3JYXc&sig=w9iQWXkv97DpyGpKSMxfmCtJB1A&hl=en&sa=X&ved=0ahUKEwiskYiY0aLJAhVLmR4KHWoNBDcQ6AEIODAF
I guess the integral part of the question was to be a big hint
it looks like I could of half used cauchy if I used the right condition... \[f(x)=cx \\ f(9999)=c \cdot 9999 \\ 3333=c \cdot 9999 \\ c=\frac{1}{3} \\ \text{ so } f(x)=\frac{1}{3} x\] and then I guess we were suppose to observe a pattern and say hey let's take the integer part of that answer
yea
\[f(x)=\text{integer part of } (\frac{1}{3} x)\]
thank you imqwerty for putting much effort into this problem with me
haha :D no problem
I'm still looking at this, I have several ideas to share however this morning has been busy, I went out to eat breakfast and planning a winter vacation with a friend in Europe so I haven't been able to chip in yet!
I used @freckles trick of changing m=2 and go this form, also I changed that 0 or 1 into a function r(n) just for my own purposes it's not really anything too special. So I start from here: \[f(k+2)-f(k)=r(k)\] From here I telescope over this sum keeping in mind that k either goes over only even or only odd numbers depending on where I start from: (If you have questions or doubts I can try to derive this, but it sort of makes more sense to me to just write it out as this) \[f(n) - f(a) = \sum_{k=0}^{(n-a-2)/2} r(2k+a)\] So I try plugging in: \[f(1982) = \sum_{k=0}^{989} r(2k+2)\] since r(n) is either 0 or 1, the best my method here does so far is puts a bound on f(1982) somewhere between 1 and 990. What remains to solve for is the general form for just the even terms of r(n). In a way though, that seems like most of the problem, all I've done is just rewrite it a little bit. Hmmm...
Join our real-time social learning platform and learn together with your friends!