Application of uncertain variables in an adaptive decision system for a complex of operations
The paper concerns a decision plant consisting of parallel operations, which are described by relational models with unknown parameters. The unknown parameters are assumed to be values of uncertain variables characterized by certainty distributions given by an expert. The solution of a decision problem and, consequently, performance of the knowledge-based decision system is very sensitive to the forms and parameters of certainty distributions. In the paper, it is shown how application of an adaptation process, consisting in step by step changing of parameters in certainty distributions based on current performance evaluation, may improve performance of a decision system under consideration. An illustrative example and results of simulations are included.
Bibliogr. 8 poz., wykr.