Optimal Decoding of Convolutional Codes in Channels with Additive Markov Noise

We formulate the conditions when the maximum of a certain function can be found by a generalized Viterbi algorithm (GVA). We prove that GVA can be used for optimal decoding of convolutional codes in channels with additive Markov noise. The complexity of the algorithm is estimated. Some simulation results supporting the efficiency of GVA are presented.