Strategies in the secretary problem (Q1329670)

From MaRDI portal
Revision as of 02:57, 5 March 2024 by Import240304020342 (talk | contribs) (Set profile property.)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
scientific article
Language Label Description Also known as
English
Strategies in the secretary problem
scientific article

    Statements

    Strategies in the secretary problem (English)
    0 references
    0 references
    3 April 1995
    0 references
    The article begins with an anecdotal, early history of one of the original, sequential, optimal stopping problems, known among other names, as the ``Secretary Problem''. A given number \(n\) of items each numbered with distinct integers (positive or negative) is drawn, sequentially, from a container. The object is to devise a ``strategy'' that provides the best chance of recognizing the largest integer as it is drawn from the container. The original proof is given that rejecting the first (approximately) \(n/e\) items and choosing the next larger one drawn (should there be one) yields a win ``\(1/e\)'' of the time. A discussion of the meaning of ``strategy'' follows, and careful proofs are given that the strategy just described is optimal in each of two classes of ``rational'' strategies, one restricted, the position strategies, and one very broad, the deterministic strategies.
    0 references
    sequential stopping
    0 references
    secretary problem
    0 references
    optimal stopping
    0 references
    position strategies
    0 references
    deterministic strategies
    0 references

    Identifiers