Channel capacity and dirty paper coding for Gaussian channels with additive and multiplicative interferences

This paper studies the capacity, as well as the performance of dirty paper coding (DPC) for a point to point power constrained fading channel given two-sided mixed state information on additive and multiplicative interferences. We show the optimality of DPC when perfect channel state information about fading variable is available to both transmitter and receiver, and provide multiplexing/dernultiplexing based strategies to achieve capacity. By employing an adaptation of Heegard's algorithm to this channel model, we also evaluate numerically both the achievable rate of the modified DPC strategy, as well as the actual channel capacity when transmitter has imperfect side information. Numerical results demonstrate significant improvements on channel capacity, as well as increased disparity between the achievable rate using DPC and channel capacity when only one-bit feedback on channel sate information of fading variable is available to encoder. The futility of knowing the additive interference at transmitter, when the multiplicative interference is only available to the receiver, is also shown numerically.