Stochastic event counter for discrete-event systems under unreliable observations
Tae-Sic Yoo, Humberto E. Garcia · 2008
This paper addresses the issues of counting the occurrence of special events in the framework of partially- observed discrete-event dynamical systems (DEDS). First, we develop a noble recursive procedure that updates active counter information state sequentially with available observations. In general, the cardinality of active counter information state is unbounded, which makes the exact recursion infeasible computationally. To overcome this difficulty, we develop an approximated recursive procedure that regulates and bounds the size of active counter information state. Using the approximated active counting information state, we give an approximated minimum mean square error (MMSE) counter. The developed algorithms are then applied to count special routing events in a material flow system.