SEQUENTIAL DECISIONS UNDER UNCERTAINTY IN THE RESEARCH-AND-DEVELOPMENT PROJECT SELECTION PROBLEM

In most cases, the R&D project selection problem is concerned with how to evaluate and identify the best subset of projects under some resou rce constraints. In this paper, we consider the sequential decision pr oblem of optimally sequencing a set of projects which must be undertak en sequentially over time. We derive the optimal ordering strategy for two special cases, 1) a series system of tasks in which the selection process is terminated as soon as one of the tasks is failed and 2) a parallel system of alternatives in which the selection process continu es until any of the alternatives is completed successfully. We further consider the precedence restriction in a series system of tasks which specifies that, for some technological or budgetary constraints, a ce rtain task must be undertaken before another. In a parallel system of alternatives, we also investigate the availability condition under whi ch a certain alternative will not be available if not selected within a certain period of time.