Optimization of Protograph LDPC Codes based on High-Level Energy Models

This paper considers the optimization of the energy consumption of LDPC decoders. For a given protograph, two models are introduced to approximate the energy consumption of a quantized Min-Sum decoder. The first model takes into account the number of operations performed by the decoder, the second model considers the number of bits that are written into memory. An optimization problem is then formulated to minimize the decoder energy consumption with respect to the protograph and the number of iterations, while satisfying a given performance criterion. Finally, an optimization method based on Differential Evolution is proposed.