Originally posted by humy
I have been mulling over some of the links I have been given and finally noticed that one clearly implies that the infinite series:
1 + 2 + 3 + 4 …
cannot possibly equal -1/12 !
http://en.wikipedia.org/wiki/Divergent_series
“...In mathematics, a divergent series is an infinite series that is not convergent, meaning that the infinite sequence of the p ...[text shortened]... rns don't tend to zero.
I now think this is probably the case. But does anyone here disagree?
For a series to be convergent the sequence of partial sums has to converge. A necessary, but insufficient, criterion for this is that lim (|a(N)|) = 0. This means that the series 1 + 2 + 3 + ... + n + ... does not converge to any finite value. This does not stop one producing a regularization scheme which assigns a finite value to the sum. We can use a formal approach where we take as axioms that if a(n), b(n) and c(n) are the nth terms of three series whose sums are A, B, and C respectively then:
c(n) = r a(n) + b(n) <=> C = rA + B.
Let's apply this to the series: S = 1 + x + x² + x³ + ...
Our axiom is enough to allow us to do the trick S = 1 + xS so we get the formal solution: S = 1 / (1 - x). We know from the other thread that the series only converges for |x| < 1. But we can analytically extend this solution to the entire plane, except x = 1, to give an expression for the sum of the series.
A case in point is S = 1 - 1 + 1 - 1 + ... which is divergent and not even conditionally convergent since lim (|a(n)|) = 1. Putting x = -1 into our formula gives x = 1/2. So we can
assign a sum to a divergent series. It is not unique as I could also do:
S = 1 - 1 + 1 - 1 + ... = (1 - 1) + (1 - 1) + ... = 0
or
S = 1 - 1 + 1 - 1 + 1 - ... = 1 + (-1 + 1) + (-1 + 1) + ... = 1
or
S = 1 - 1 + 1 - 1 + 1 - ... = 1 + (-1 + 1 - 1 + 1 + ...) = 1 + (1 - 1 + 1 - 1 + ...) = 1 + 1 + (-1 + 1) + (-1 + 1) + ... = 2
Since we can also sum the remainder of the series to ½ we can get any half integer result. So we have to be very careful when assigning a value to a series.
In the same way S = 1 + 2 + 3 + 4 + ... is a special case of
S(s) = 1 + 2^-s + 3^-s + 4^-s + ... = zeta(s).
By analytic continuation the
divergent series can be
assigned the value zeta(-1) = -1/12. The harmonic series (which is less obviously divergent) cannot be assigned a finite value using this regulation scheme.
Look up the Wikipedia page on Casimir energy if you want to see an example of this kind of regulation scheme being used to calculate a physically meaningful quantity.