Up to the present dynamic programming has been regarded as a general
type of approach to problem solving, essentially based on decomposition
of the given problem into a sequence of smaller subproblems.
It has permitted the solution of a remarkable number of problems (deterministic
and stochastic, continuous and discrete, constrained and unconstrained,
static and dynamic, and so on). In general, however, each
individual situation has required derivation from the general philosophy
of an ad hoc algorithm.