Optimization of Quantum Information Processing Maximizing Mutual Information

A model of quantum noisy channel with input encoding by a classical random vector is described. An equation of optimality is derived to determine a complete set of wave functions describing quantum decodings based on quasi-measurements maximizing the classical amount of transmitted information. A solution of this equation is found for the Gaussian multimode case with input Gaussian distribution. It is described by the overcomplete family of coherent vectors describing an optimal quasimeasurement of the canonical annihilation amplitudes in the output Hilbert space. It is found that the optimal decoding in this case realizes the maximum amount I=Spln[1+S/(N+1)] of the classical information as transmitted via the classical Gaussian channel with the effective noise covariance matrix N+I. A physical realization of optimal quasi-measurement based on an indirect (heterodyne) observation of the canonical operators is suggested.