Quantifying the computational efficacy of nanocomputing channels

Information-theoretic computational efficacy measures are introduced and applied to the analysis of interacting effects of noise and structural randomness in nanocomputing channels. Two complementary measures–the computational fidelity and representational faithfulness–are shown together to quantify the efficacy with which a particular noisy computing channel implements a specified logical transformation. The statistics of these two measures, evaluated for an ensemble of nanocomputing channels with random sample-to-sample structural variations, allows quantitative characterization of the interacting effects of transient noise and structural randomness in the channel ensemble. Application of these efficacy measures is illustrated for artificial nanocomputing channels–noisy quantum-dot cellular automata arrays with random defects–although they can be used to quantify the computational efficacy of a wide variety of artificial or natural nanoscale networks with computational capabilities.