Zero risks in one-criteria tasks

Keywords

pure strategy
criterion
uncertainty
Savage risk functions
minimax regret principle
mixed strategy
risk

How to Cite

Жуковский, В. И., & Горбатов, А. С. (2015). Zero risks in one-criteria tasks. Risk Managmenent, 2. https://riskmanagementjournal.ru/index.php/rmj/article/view/226

Abstract

American mathematician, economist and statistician Leonard Jimmie Savage (1951) suggested the minimax regret principle (MRP), which along with the maximin principle plays an important role in accepting the guaranteed solutions in one-criterion problems with uncertainty («game with nature»). In MRP Savage used the regret function, which later became known as «Savage risk function». Its value defines the risk, associated with the action (strategy) chosen by the DM (decision maker). DM is trying to make such a risk as little as possible, preferably down to zero. In this paper we establish the existence of a mixed strategy, that guarantees «the best», zero risk, under the «usual» for mathematical game theory restrictions.