This paper discusses dynamic methods for solving a class of multi-project s
cheduling problems in which rates of job performances are controllable and
resources such as money, energy or manpower per time unit, are renewable an
d continuously divisible. The objective is to complete the projects as clos
e to the common due date as possible. Two different ways of imposing sequen
tial precedence relations between project jobs are explored by formulating
two dynamic models and studying their relationships on the optimal solution
. Efficient time-decomposition algorithms for finding either globally optim
al schedules or lower bound guided near-optimal solutions are suggested and
computationally tested.