12-1-09 Find summation of this series...

PHEW.... finally here.....

f(0)=N g(0)=0

g(n+1)=2{f(n)/2+g(n)/2} (fractional part)

f(n+1) = [f(n)/2+g(n)/2] (greatest integer..)

where n is integer

find f(1)+f(2)+f(3)...... infinity...

N is an integer :)

39 Answers

62
Lokesh Verma ·

okie...

magar kaise??!!!

62
Lokesh Verma ·

haha.. this is what i wanted to hear.. :) :D

See to be true this is not a jee problem..... But it will add a new dimension to your thought process

1
Philip Calvert ·

[11][11][11][11][11][11][11][11][11][11][11][11][11][11][11][11][11][11][11][11][11][11][11][11][11][11][11][11][11][11][11][11][11][11][11][11][11][11][11][11][11][11][11][11][11][11][11][11][11][11][11][11][11][11][11][11][11][11][11][11][11][11][11][11][11][11][11][11][11][11][11][11][11][11][11][11][11][11][11][11]
I agree with sky and that is why u have deleted that part about it not being that bad bhaiyya

62
Lokesh Verma ·

no it is not that bad :P

but bcos to think of the question took me half an hour... so i thought that i rather remove the phrase :D

1
Philip Calvert ·

ok got a sly trick[4]
will post something after 10 mins[4]

1
Philip Calvert ·

thank goodnes u forgot about writing that negative mark thing about the explanation part [1]

1
Philip Calvert ·

that should not be the way to see it i know but is the answer correct

62
Lokesh Verma ·

f(0)=N g(0)=0

g(n+1)=2{f(n)/2+g(n)/2} (fractional part)

f(n+1) = [f(n)/2+g(n)/2] (greatest integer..)

where n is integer

N=3

f(0)=3

g(1)=2{3/2}=1
f(1)=[3/2]=1
g(2)=2{1/2+1/2} =0
f(2)=[2/2]=1
f(3)=0
f(4)=0...

hence f(1)+f(2) = 2

does this give some hint?

1
Philip Calvert ·

sorry so stupid of me

1
Philip Calvert ·

ans looks to be N-1
yes N-1

edited
P.S jus read chatbox i hadn't read when i posted this :)

1
skygirl ·

HUH!!

THIS IS BAD!

VERY BAD!

9
Celestine preetham ·

λproved using induction

F(N) = f1 + f2 ............. for N

then F(N+1)= F(N) +1

this is proved by taking N=4λ + (1or 2 or3or4)

and similarly representing N+1 in same form after 3 steps( f1 ,f2,f3) ull find that theyre sum differs by one but their parameters ie f3,g3 bcom same
so rest of sum is identical and induction finished

also F(1) is 0 so F(N)=N-1

1357
Manish Shankar ·

hint take

N=a0+2a1+4a2..... + 2kak

Now try to solve

where each ai is either 0 or 1

before that you have to justify that the above can be done!

62
Lokesh Verma ·

celestirne how is that!! :O

I din get how you proved the induction hypothesis!

9
Celestine preetham ·

manish sir i already gave that hint of rep N as a binary form in my previous post but then i found a simpler not so confusing method so i left of that binary form method.

see take the foll values with corresponding f,g after each step

4λ+1 4λ+2 4λ+3 4λ+4 4λ+5

2λ,1 2λ+1,0 2λ+1,1 2λ+2,0 2λ+2,1

λ,1 λ,1 λ+1,0 λ+1,0 λ+1,1

case λ=2μ

μ,1 μ,1 μ,1 μ,1 μ+1,0
as f + g is same in all cases remaining values are identical

similarly take for λ = 2μ+1 also

now wen u add up till the point were they dont repeat u find succesive F(N) differ by 1

this was the proof i was talkin abt

but i feel u ve a more beautiful proof involving binary nos that im currently not able to think abt due to my ongoing chem exams ;(

9
Celestine preetham ·

pls post ur sol

62
Lokesh Verma ·

I will give a veyrr very different proof...

There is a knock out tennis match of N player..

What is the number of matches required to chose the winner.

At each level, the player who does not play goes to the next level...

Then the expression is the number of matches in each level....

The number of matches is n-1 because at each game, 1 player is eliminated :)

Dont kill me for this solution... I told that this one is the toughest (may be stupid) question u will ever come across :P)

9
Celestine preetham ·

great q nish
im sure this q is of inmo level but then its hard to thnk in ur trms cos u already know the sol ;)
ne way i solved it usin long method pls give proof using binary nos

62
Lokesh Verma ·

haha..

well u know what .. this question came to me because i had this question.. in permutation.. and the series that i got was the one i gave...

And i was just like in no position to know what the sum would be..

then i realised that the answer is N-1!!

So i made this question from there! :O

The binary method.. I am not very sure... :(

I will have to think for that... (My gut feeling says that it can be done by that!)

May be one way could be prove for any 2n (which sis straightforward) Then prove for sum of numbers.... (But i am not sure if that will work!)

9
Celestine preetham ·

even my gut feelin was that but i was unable to proceed further so changed to another method

62
Lokesh Verma ·

yup i have... let me fix it ... before more guys start to think.... too hard.. :(

62
Lokesh Verma ·

yes

9
Celestine preetham ·

k

series = n( 1/2 + 1/4 ...) + (1-1/2) +(1-1/4)+(1-1/8).......

= n + ∞ - 1

;)
r u sure the q is rite or am i rong

9
Celestine preetham ·

oh sry i din see that as greatest int fn neway will try ,the abv mite be a clue

62
Lokesh Verma ·

a small mistake.. see the new question... :)

1
Philip Calvert ·

[12][12]

62
Lokesh Verma ·

let me give an example....

for 5

f(5)=[(5+1)/2]=3

ff(5)=[3/2]=1

fff(5) = 0

sum will be 3+1 = 4

1
Philip Calvert ·

i think u have written something wrong [12]

9
Celestine preetham ·

sry nish q rong
fff5 = 1 and this ll continue forever

1
Philip Calvert ·

how ff(5) = [3/2] shud be [4/2]

Your Answer

Close [X]