Abstract: A method evaluates and optimizes an error-correcting code to be transmitted through a noisy channel and to be decoded by an iterative message-passing decoder. The error-correcting code is represented by a parity check matrix which is modeled as a bipartite graph having variable nodes and check nodes. A set of message passing rules is provided for the decoder. The decoder is analyzed to obtain a set of density evolution rules including operators and operands which are then transformed to projective operators and projected operands to generate a set of projective message passing rules. The projective message passing rules are applied iteratively to the error-correcting code modeled by the bipartite graph until a termination condition is reached. Error rates of selected bits of the error-correcting code are then determined by evaluating the corresponding operands. The error rates can be passed to an optimizer to optimize the error-correcting code.

Abstract: A method evaluates and optimizes an error-correcting code to be transmitted through a noisy channel and to be decoded by an iterative message-passing decoder. The error-correcting code is represented by a parity check matrix which is modeled as a bipartite graph having variable nodes and check nodes. A set of message passing rules is provided for the decoder. The decoder is analyzed to obtain a set of density evolution rules including operators and operands which are then transformed to projective operators and projected operands to generate a set of projective message passing rules. The projective message passing rules are applied iteratively to the error-correcting code modeled by the bipartite graph until a termination condition is reached. Error rates of selected bits of the error-correcting code are then determined by evaluating the corresponding operands. The error rates can be passed to an optimizer to optimize the error-correcting code.