Hierarchical control problems [1, 2] are, as a rule, investigated with
in the framework of the theory of differential games [3], though they
are intrinsically closer to the classical control problems than the an
tagonistic differential games, which have been intensively developing
in recent times. The classification of the problems related to the beh
avior of dynamic objects under goal-oriented interactions into control
and game problems is rather artificial. In this classification, class
ical control problems are put at one endpoint, which are characterized
by a unified goal for all (possibly, distinct) participants of the co
ntrol process. Antagonistic differential games are put at the other en
dpoint, in which the participants in one of two groups regard each oth
er as opponents or malevolant disturbances. In real conditions, the ma
jority of interesting problems occupy an intermediate position, which
is determined by the degree and form of the interactions (behavior pri
nciples) of the participants with opposing interests. The hierarchical
problems among control problems are distinguished by the behavior of
the participants. In these problems, the participants do not have a co
mmon goal: each pursues his own goal, which is not consistent with the
interests of others, but they all observe a strict (hierarchical) act
ion in the control process [4]. In this paper, we develop an approach
to synthesizing optimal controls for linear hierarchical control probl
ems.