OPTIMIZATION OF THE AUTOMATIC SEARCH PROCESS UNDER THE EFFECT OF RANDOM PERTURBATIONS (OPTIMIZATSIYA PROTSESSA AVTOMATICHESKOGO POISKA PRI VOZDEISTVEII SLUCHAINYKH VOZMUSHCHENII),
Abstract
It is pointed out that a random noise at the measuring-element input of an extremal system increases the time of extremum search. The probability of a false searching step can be reduced by a statistical averaging of several controlled-hunting steps, which makes these steps longer and, therefore, increases the total search time. It is noted that the Wald method of statistical sequential analysis is the optimum method with respect to the acceptable number of hunting steps consistent with a sufficiently short search time. This method investigates the random process by observing it and by setting up hypotheses about the nature of the process (for example, choosing between two solutions of two alternative situations). The method is peculiar in that it involves (a) a multistage statistical experiment and (b) the decision to desist or continue the experiment depending on a random previous step.
Document Details
- Document Type
- Technical Report
- Publication Date
- Nov 06, 1967
- Accession Number
- AD0675275
Entities
People
- N. E. Madzharov
Organizations
- National Air and Space Intelligence Center