Fuzzy successive averaging method in multistage optimization systems

In this paper a new method which extends the classical method of successive averages used in flow assignment problems to the stochastic flows based on fuzzy aggregation is developed. The link capacities assignment is of minimum risk type and the measurements on the traffic patterns flowing through these links are time limited (that is, the congestion processes are predicted and tracked using a minimum quantity of information about the system history). A regression model for identification of the fuzzy parameters is considered. The application of this method to multistage decision making in stochastic flows assignment in digital communication networks is applied for the case of link capacities with Gaussian modulation