Article ID: | iaor19931451 |
Country: | United Kingdom |
Volume: | 31 |
Start Page Number: | 16 |
End Page Number: | 28 |
Publication Date: | Dec 1991 |
Journal: | USSR Computational Mathematics and Mathematical Physics |
Authors: | Baranov V.V. |
Keywords: | computational analysis |
The computation of an optimal strategy in the sense of the expected utility criterion in a discrete-time optimal stochastic control model is considered. An optimality principle is formulated, which defines the necessary and sufficient conditions of optimality of a stationary strategy. A general successive-approximation optimization scheme is developed. This scheme produces a whole family of new optimization methods.