I think this paper may help:

M. Bellare, T. Krovetz and P. Rogaway (1998),
["Luby-Rackoff backwards: Increasing security by making block ciphers non-invertible"](http://www.cs.ucdavis.edu/~rogaway/papers/p2f-abstract.html), _Advances in Cryptology - EUROCRYPT '98_, Lecture Notes in Computer Science, Vol. 1403.

> "**Abstract:** We argue that the invertibility of a block cipher can reduce the security of schemes that use it, and a better starting point for scheme design is the non-invertible analog of a block cipher, that is, a pseudorandom function (PRF). Since a block cipher may be viewed as a pseudorandom permutation, we are led to investigate the reverse of the problem studied by Luby and Rackoff, and ask, how can one transform a PRP into a PRF in as security-preserving a way as possible? The solution we propose is "data-dependent re-keying." As an illustrative special case, let $E: \{0,1\}^n \times \{0,1\}^n \to \{0,1\}^n$ be the block cipher. Then we can construct the PRF $F$ from the PRP $E$ by setting $F(k,x) = E(E(k,x),x)$. We generalize this to allow for arbitrary block and key lengths, and to improve efficiency. We prove strong quantitative bounds on the value of data-dependent re-keying in the Shannon model of an ideal cipher, and take some initial steps towards an analysis in the standard model."

[A PDF of the paper is available here.](http://www.cs.ucdavis.edu/~rogaway/papers/p2f.pdf) (The PDF link on the abstract page is wrong, it leads to the PS version.)