AN OPTIMUM DISCRETE SPACE SEQUENTIAL SEARCH PROCEDURE WHICH CONSIDERS FALSE ALARM AND FALSE DISMISSAL INSTRUMENT ERRORS.
SOUTHERN METHODIST UNIV DALLAS TEX DEPT OF STATISTICS
Pagination or Media Count:
An optimum sequential search policy is developed for locating an object in a discrete search space assuming a search instrument having both false alarm and false dismissal errors. The search policy obtained specifies that the experimenter always takes his next observation at the potential location which has the largest a posteriori probability of being the true location as computed from the prior probability distribution and previous experimental evidence via Bayes Rule. The results were derived under the following assumptions. A single object is assumed to be located in 1 of N locations and to remain stationary during the search process. The experimenter is presumed to have an initial discrete a priori probability function describing the location of the object. Also available is a single search instrument which the experimenter can employ to examine each of the potential locations. Only one location can be observed at any one time, however. The instrument is assumed to perform a simple, fixed sample size, hypothesis test having known conditional error probabilities of both Type I and Type II. The preliminary decision made by the instrument is thus a noisy decision which is only statistically related to the presence or absence of the object at the location in question. A communication system search example is also considered. Author
- Operations Research