- #1

dyh

- 11

- 0

I got f(n)=n^2.

I can not find anymore solution except this.

I just wonder there are some more solutions about this problem or not.

I think there are more, but I don't know how to get to them.

I want to hear your opinion.

Thanks

You are using an out of date browser. It may not display this or other websites correctly.

You should upgrade or use an alternative browser.

You should upgrade or use an alternative browser.

- Thread starter dyh
- Start date

- #1

dyh

- 11

- 0

I got f(n)=n^2.

I can not find anymore solution except this.

I just wonder there are some more solutions about this problem or not.

I think there are more, but I don't know how to get to them.

I want to hear your opinion.

Thanks

- #2

Bacle2

Science Advisor

- 1,089

- 10

Have you considered general methods for solving recursions:

http://en.wikipedia.org/wiki/Recurrence_relation ?

- #3

dyh

- 11

- 0

Thanks

I know how to solve the general methods for solving recursions.

But I think this is some special case of it because there is some contradiction in particular solution.

I.e. when I put f(n)= c (constant) for non-negative integer n. then

I can get

c=(1/2)c+(1/2)c-1

so it looks like 0=-1

So, I need to try "cn" but there is also contradiction in this case similar to previous case.

then I need to try "cn^2". In this case, if I choose c as 1 then it can be a solution.

So, I just would like to know this is the only solution or not.

If not, I want to know how to get other solutions.

- #4

dyh

- 11

- 0

oh.. I got some other general solution form

f(n) = a+bn+n^2, a,b are constant

But I still would like to know there would be more solutions or not. hehe

- #5

Bacle2

Science Advisor

- 1,089

- 10

I know how to solve the general methods for solving recursions.

But I think this is some special case of it because there is some contradiction in particular solution.

I.e. when I put f(n)= c (constant) for non-negative integer n. then

I can get

c=(1/2)c+(1/2)c-1

so it looks like 0=-1

So, I need to try "cn" but there is also contradiction in this case similar to previous case.

then I need to try "cn^2". In this case, if I choose c as 1 then it can be a solution.

So, I just would like to know this is the only solution or not.

If not, I want to know how to get other solutions.

Sorry, I don't see it. Don't you have to choose the first two terms to find out the

value of f(n+1). Then, if you choose f(n)=f(n-1)=c , you get f(n+1)=2+c;

I don't see any contradiction.

- #6

dyh

- 11

- 0

In my opinion,

if I choose f(n)=f(n-1)=c for any non-negative integer n, then f(n+1) would be "c" too

- #7

Bacle2

Science Advisor

- 1,089

- 10

O.K, let's see:

f(n)=(1/2)f(n+1)+ (1/2)f(n-1)-1

Set f(n)=c=f(n-1) . Then,

c= (1/2)f(n+1)+c/2-1 , so:

1+ c- c/2 = (1/2)f(n+1), so :

1+c/2= (2+c)/2=(1/2)f(n+1) , so f(n+1)=2+c .

That's what I get.

Share: