Channel power allocation factors computing in MIMO wireless networks with improved water pouring algorithm

In this paper we consider MIMO system with Mt transmitting and Mr receiving antennas, when channel state information (CSI) is known on the transmitter side. In this case, the optimum multiple antenna transmission consists of allocating the transmitted power for each virtual sub-channel related to the corresponding eigenvalue of the propagation channel matrix. The optimum power allocation is computed using the water pouring algorithm (WPA). However, on-line implementation of the algorithm requires serious computational work, which is O(M_t ²). We propose a modification to WPA that reduces the computational complexity to O(M_t).