Sequential strategies in dual control problems

✍ Scribed by Richard M. Cyert; Morris H. Degroot

• Publisher: Springer US
• Year: 1977
• Tongue: English
• Weight: 844 KB
• Volume: 8
• Category: Article
• ISSN: 0040-5833
• DOI: 10.1007/bf00133410

✦ Synopsis

Sequential decision problems are studied in which there are two decisionmakers, or players, who are trying to control the same stochastic system in order to minimize their own individual expected losses. The standard linear model for the stochastic control process and standard quadratic loss functions are assumed. The players take turns choosing the value of the control variable, and each player has his own sequence of targets and his own loss function. Optimal pure competitive strategies are derived by backward induction, and limiting strategies and equilibria are determined. Myopic strategies are introduced and shown to yield smaller losses for each player than the pure competitive strategies. Coefficients of cooperation are then defined, and various cooperative sequential strategies based on them are shown to be mutually beneficial to both players.