ANOTHER OPTIMAL STOPPING PROBLEM,
RAND CORP SANTA MONICA CALIF
Pagination or Media Count:
Randolph has discussed the following problem suppose an experimenter takes observations X1, X2, ... sequentially from a given population. Each observation costs a fixed amount c. After each observation the experimenter can either stop and take as his reward the maximum of the Xi so far observed, or he can take another observation. The problem is to discover what policy will maximize net return the difference between reward and costs. The purpose of this note is to discuss a problem that leads to an optimal stopping rule which terminates the experiment whether the largest observed X is small or large.
- Operations Research