Utilization of Distributed Computing in Exhaustive Search for Optimal Block Code

This paper presents an implementation of distributed computing to search for the optimal block error control code (covering both linear and nonlinear codes) for binary symmetric channels. The optimality is determined by comparing the decoding error probabilities when maximum likelihood decoders are used under the assumption that the codewords are equally likely. Several pre-processing steps are proposed to reduce the size of the search space. The search gives distance distributions which correspond to closed-form expressions for the minimum decoding error probability. Distributed computing is utilized to further reduce the search time. The proposed implementation can be extended to an arbitrary number of cloud servers. The amount of time saved by the search space reduction and distributed computing is also demonstrated.