A Model and Algorithm for Analysis of Discrete Event Systems with Structural Changes

The paper developes the Combined Simulation and Perturbation Analysis (CSPA) algorithm. CSPA integrates perturbation analysis (PA) techniques and simulation to generate performance information about a new system (the perturbed system) usig information about a preceding system (the nominal system). This will aid system designers in analyzing and comparing the performance of systems which may differ structurally parametrically, or both. The discrete event system (DES) model (based on the GSMP model) and the framework for comparing different discrete event systems are presented. The CSPA algorithm which leverages system similarities when simulating new systems is discussed and its correctness is proven. The CSPA algorithm's contribution will be discussed, including an example with results.