Applications of the theory of optimal design of experiments to radial velocity (RV) planet search surveys are considered. Different optimality criteria are discussed, basing on the Fisher, Shannon and Kullback-Leibler informations. Algorithms of optimal scheduling of RV observations for two important practical problems are considered. The first problem is finding the time for future observations to yield the maximum improvement of the precision of exoplanetary orbital parameters and masses. The second problem is finding the most favourable time for distinguishing alternative orbital fits (the scheduling of discriminating observations). These methods of optimal planning are demonstrated to be potentially efficient for multiplanet extrasolar systems, in particular for resonant ones. In these cases, the optimal dates of observations are often concentrated in quite narrow time segments.