# algorithm for generating interesting normal symmetric continuous functions

Does any one know of, or can think of an algorithm which generates arbitrarily many symmetric normally distributed continuous functions? And when I say interesting I mean more complex distributions then an offset bell curve.

-
I don't understand what you are asking for. Taking a random number generator and getting a sequence of means and variances, and then considering the Normal distribution with those parameters seems to satisfy your conditions but hardly is what you are looking for. Can you clarify a little bit? –  gt6989b Oct 28 '13 at 19:30
@gt6989b I'm trying to generate continuous functions, so that I can integrate them. They need to be symmetric and normalize-able. Or better yet the algorithm would make them normal to begin with. –  Loourr Oct 28 '13 at 19:40
So why is my suggestion not good (i.e. generate random $\{\mu_i, \sigma_i\}$ and each pair implies a Normal distribution provided $\sigma_i > 0$)? –  gt6989b Oct 28 '13 at 19:58
@gt6989b it's not bad, I'm just not sure how to turn that into a continuous function –  Loourr Oct 28 '13 at 20:34

Clarifying our discussion in the comments.

Consider generating a sequence of real pairs $\left(\mu_k, \sigma_k\right)_{k=1}^N$ with $\sigma_k > 0$ for each $k$.

Then, for each $k$, consider the pdf of the $\mathcal{N} \left(\mu_k, \sigma_k \right)$ distribution, which is given by

$$f(x) = \frac{\exp \left( -\frac{1}{2} \left(\frac{x-\mu_k}{\sigma_k}\right)^2 \right)} {\sigma_k \sqrt{2\pi}}.$$

-