#1 - e. Note stirling's approximation to n! is √(2πn)(n/e)^n. Taking the nth root of this, we get (2π)^(1/(2n)) * √(n^(1/n) * n/e. Claerly, as n→∞, (2π)^(1/(2n))→1. Note that n^(1/n)=e^(ln n/n). The limit of ln n/n as n→∞ is 0, and e^0 is 1, and √1 is 1, so as n→∞, √(n^(1/n) also approaches 1. Thus as n→∞, n!^(1/n) →n/e, which mean that n/(n!^(1/n)) → e.
#2 - Actually, yes, though it's not because of the infinite degree polynomial, but rather because of Euler's identity, e^(ix) = cos x + i sin x. The soutions are of the form x=0+2iπk, where k is any integer. Thus, you might have x=0 (trivial, since anything to the 0=1), but also x=2iπ, x=4iπ, x=-3764iπ, etc.
2006-07-09 00:04:19
·
answer #1
·
answered by Pascal 7
·
1⤊
0⤋
I wanted to remark on problem 2 a bit. Note that exp(x) can be written as an "infinity degree" power series, yet the equation exp(x)=0 has no solution (even complex ones). It's an interesting exercise (with a computer) to see what happens to the zeroes of the finite MacLaurin polynomials of exp(x), as the degrees go up.
So you'd start with 1,1+x,1+x+x^2/2,1+x+x^2/2+x^3/3!, etc., which have 0, 1, 2, 3, etc. roots.
2006-07-09 12:17:25
·
answer #2
·
answered by Steven S 3
·
0⤊
0⤋
2. Yes. solutions are :
s = lim A / (B)^n ; A, |B| > 1, from R
n->00
for instance s = (+/-) 0.0000000000000000000000..........
or 1 - 0.99999999999999 ....
2006-07-09 06:31:20
·
answer #3
·
answered by gjmb1960 7
·
0⤊
0⤋
1-answer is 1
2-answer is 1^1,y^0
2006-07-09 05:37:55
·
answer #4
·
answered by dawit f 1
·
0⤊
0⤋
yes
2006-07-09 05:32:32
·
answer #5
·
answered by fluorescent`pen 3
·
0⤊
0⤋
something i dont understand
2006-07-09 08:29:31
·
answer #6
·
answered by Prakash 4
·
0⤊
0⤋