An approach to formalize structural decomposition and aggregation for stochastic reward net models

The paper presents an approach to formalize decomposition and aggregation of stochastic reward nets based on their structure. The set of places and transitions are automatically partitioned. Sets of these partitions are aggregated. Partial and full aggregates are built for an iterative scheme to calculate approximately stochastic rewards. In other words, approximate performance evaluation is fully automated for any live, bounded and reversible stochastic reward net as far as at least one aggregable partition is formed. The technique is applied to the analysis of a flexible manufacturing system.