Optimal Search for the Wake of a Moving Target When Searcher Motion is Constrained.

摘要

A method for determining the optimal or near-optimal search path for the wake of a moving target when the searcher's motion is constrained is presented. The problem uses a Markov motion model in discrete time and space for the target and assumes that the searcher is constrained to move only from the currently occupied cell j to a specified set of 'neighbor cells', I(j). First, a discussion of the complexity of the problem is presented. Next, an extension of T.J. Stewart's constrained searcher algorithm is given. Stewart's algorithm uses S.S. Brown's unconstrained searcher algorithm to calculate bounds on the probability of nondetection. An extension of Brown's algorithm to allow the use of a wake detector is also given. Several alternatives to both algorithms are offered and compared. Finally, some further extensions to the algorithms are suggested. (Author)