1: \begin{abstract}

2: Density evolution is one of the most powerful analytical tools for low-density

3: parity-check (LDPC) codes and graph codes with message passing decoding

4: algorithms. With channel symmetry as one of its fundamental assumptions,

5: density evolution (DE) has been widely and successfully applied to different

6: channels, including binary erasure channels, binary symmetric channels, binary

7: additive white Gaussian noise channels, etc. This paper generalizes density

8: evolution for {\it non-symmetric} memoryless channels, which in turn broadens

9: the applications to general memoryless channels, e.g.\ z-channels, composite

10: white Gaussian noise channels, etc. The central theorem underpinning this

11: generalization is the convergence to perfect projection for any fixed size

12: supporting tree. A new iterative formula of the same complexity is then

13: presented and the necessary theorems for the performance concentration theorems

14: are developed. Several properties of the new density evolution method are

15: explored, including stability results for general asymmetric memoryless

16: channels. Simulations, code optimizations, and possible new applications

17: suggested by this new density evolution method are also provided. This result

18: is also used to prove the typicality of linear LDPC codes among the coset code

19: ensemble when the minimum check node degree is sufficiently large. It is shown

20: that the convergence to perfect projection is essential to the belief

21: propagation algorithm even when only symmetric channels are considered. Hence

22: the proof of the convergence to perfect projection serves also as a completion

23: of the theory of classical density evolution for symmetric memoryless channels.

24: \end{abstract}

25: