assistance needed with infinite series proof



 Science > Physics > assistance needed with infinite series proof

LINK TO THIS PAGE  


rating :  0   |  0


  Page 1 of 1

1

 
Topic: Science > Physics
User: ""
Date: 04 Jun 2005 03:54:48 PM
Object: assistance needed with infinite series proof
First, let me describe a simpler problem I can already prove:
Consider an infinite sequence whose elements are {c(1), c(2), c(3),
....} such that
c(i+1) = a c(i)
for all i=1,2,...,infinity and some fixed real number "a". Consider
the infinite series
S = sum(i=1,infinity) { c(i) / R^i },
where R is some fixed positive number greater than unity. "R^i" means
"R to the power i."
Then Proposition #1 is:
For any pair {a,R} such that the series S converges to a finite real
number, then
S = B c(1)
for some *POSITIVE* real number B. (In other words, convergence of S
implies B >= 0.)
Proof: The proof of this is easy (just consider the geometric
series).
So far, so good. what I want to do is to generalize Proposition #1 to
the following:
Problem: Consider again an infinite series {c(1), c(2), c(3), ...} but
this time suppose
c(i+2) = x c(i+1) + y c(i)
for all i=1,2,...,infinity and some fixed real numbers {x,y}. Consider
again the infinite series
S = sum(i=1,infinity) { c(i) / R^i },
where R is some fixed positive number greater than unity. Show that:
For any triplet {x,y,R} such that S converges to a finite real
number, then
S = E c(2) + F c(1)
for some positive real number E, where {E,F} are both real numbers. In
other words,
show that convergence of S means that E must be non-negative.
[Note: I suspect B must be positive for any {x,y,R} that lets S
converge. However, if this is not generally
true, then my question would be: What are sufficient and necessary
conditions on {x,y} that implies
B must be positive whenever S converges.]
Thanks in advance!
.

User: ""

Title: Re: assistance needed with infinite series proof 04 Jun 2005 04:26:25 PM
Oh yeah, I already know that I could just try to crank out the solution
of my problem by brute force (e.g., solve explicitly for c(i) and then
summing the series). However, this does not really help me, because I
eventually want to be able to generalize the proposition further to
more complicated c(i) sequences.
Thus, I am looking for a deeper, broader insight into what general
conditions guarantee that E >=0.
Is there some pre-existing theorem about geometric series of AR(N)
sequences that will help?
Thanks again!
.

User: "Jannick Asmus"

Title: Re: assistance needed with infinite series proof 04 Jun 2005 04:09:03 PM
On 04.06.2005 22:54,
wrote:

First, let me describe a simpler problem I can already prove:

Consider an infinite sequence whose elements are {c(1), c(2), c(3),
....} such that
c(i+1) = a c(i)
for all i=1,2,...,infinity and some fixed real number "a". Consider
the infinite series
S = sum(i=1,infinity) { c(i) / R^i },
where R is some fixed positive number greater than unity. "R^i" means
"R to the power i."

Then Proposition #1 is:
For any pair {a,R} such that the series S converges to a finite real
number, then
S = B c(1)
for some *POSITIVE* real number B. (In other words, convergence of S
implies B >= 0.)
Proof: The proof of this is easy (just consider the geometric
series).

So far, so good. what I want to do is to generalize Proposition #1 to
the following:

Problem: Consider again an infinite series {c(1), c(2), c(3), ...} but
this time suppose
c(i+2) = x c(i+1) + y c(i)
for all i=1,2,...,infinity and some fixed real numbers {x,y}. Consider
again the infinite series
S = sum(i=1,infinity) { c(i) / R^i },
where R is some fixed positive number greater than unity. Show that:
For any triplet {x,y,R} such that S converges to a finite real
number, then
S = E c(2) + F c(1)
for some positive real number E, where {E,F} are both real numbers. In
other words,
show that convergence of S means that E must be non-negative.

[Note: I suspect B must be positive for any {x,y,R} that lets S
converge. However, if this is not generally
true, then my question would be: What are sufficient and necessary
conditions on {x,y} that implies
B must be positive whenever S converges.]

Thanks in advance!

Perhaps the following idea can be half-way of a solution of your problem:
There is a general solution of the two-step-iteration - like the
solution as a geometric series in the one-step-case: Try the Ansatz
c(i) = a * exp(k*i) + b * exp(l*i).
Then you'll find out that the constants k,l fulfill the eigen-equation
z^2=x*z +y. Solve for the two solutions of this equation and calibrate
a,b to c(0), c(1).
Cheers,
J.
.


  Page 1 of 1

1

 


Related Articles
 

NEWER

pg.1612     pg.1232     pg.940     pg.716     pg.544     pg.412     pg.311     pg.234     pg.175     pg.130     pg.96     pg.70     pg.50     pg.35     pg.24     pg.16     pg.10     pg.6     pg.3     pg.1

OLDER