Multifrequency synthesis algorithm based on the generalized maximum

We propose the multifrequency synthesis (MFS) algorithm with the spectral correction of frequency-dependent source brightness distribution based on the maximum entropy method. In order to take into account the spectral terms of nth order in the Taylor expansion for the frequency-dependent brightness distribution, we use a generalized form of the maximum entropy method. This is suitable for the reconstruction of not only positive-definite functions, but also sign-variable functions. With the proposed algorithm, we aim to produce both an improved total intensity image and a two-dimensional spectral index distribution over the source. We also consider the problem of the frequency-dependent variation of the radio-core positions of self-absorbed active galactic nuclei, which should be taken into account in a correct MFS. The proposed MFS algorithm has first been tested on simulated data and then applied to the four-frequency synthesis imaging of the radio source 0954+658 using Very Large Baseline Array observational data obtained quasi-simultaneously at 5, 8, 15 an