Self-adaptive simulation time for sensitivity analysis of a stochastic computational model

Self-adaptability is the property of a system through which it adjusts its behavior in response to the perception of the environment and the system itself. The engineering of self-adaptive systems is a very active research field; more specifically, several studies are focused on how to implement self-adaptive mechanisms able to deal with variability of system´s properties while preserving the attainment of the goals that have to be achieved. In this work, we describe how sensitivity analysis, a very important technique to study the behavior of computational models, can be viewed as an analytical system that has to deal with the variability of model dynamics when different model configurations are exploited. In particular, through the evaluation of a specific case study, it is shown how the development of an appropriate self-adaptive mechanism, based on a feedback control loop, enables the capability to deal with variable model dynamics in a steady state sensitivity analysis of stochastic simulations.