The "Paralysis by analysis" problem in resource allocation control with application to SDI

The "optimal" real-time allocation of resources in a typical SDI construct involves a number of difficult decision processes involving scarce resources, conflicting requirements, operational uncertainties and a time-stressed environment. One of the most interesting variants in this class of control problems is the problem of optimally deciding scarce resource allocations when the optimal decision process itself consumes the same resource that is to be allocated. In other words, the more analysis one performs, in an effort to compute the "optimal" allocation decision, the less resource one has to allocate. In fact, if one expends too much effort in computing the optimal allocation decision there may be no resources left to allocate! This situation is a control theory manifestation of the famous "paralysis by analysis" condition introduced in the political writings of former U.S. President Richard M. Nixon. In this paper, we first examine the general control-theoretic features of the "paralysis by analysis" problem in resource allocation control and derive two fundamental results (the "universal critical curve" and the "system benefit curve") which govern all such problems. Then, we describe some particular resource allocation situations, within SDI, where such problems naturally arise.