As pointed out in section 2.4, multi-decision problems are represented naturally by trees in which a node x represents the set of (component) decisions actually made, and the branches from x represent the possible x‘choices of action at x’. The problem of finding an optimal composite decision is equivalent to the problem of finding an optimal path from the root to a terminal node in a search tree (see section 2.4), and one technique for achieving this aim is that of discrete dynamic programming as shown in chapter 2. In this chapter an alternative technique, called branch-and-bound, is introduced and described.
Unable to display preview. Download preview PDF.