Power and temperature model in the presence of loose information, state-space exploration
Thursday 8 February 2018
Previous work showed how to co-simulate a functional model with a power and temperature solver. However, the previous approaches assumed known values for each parameter, which is rarely the case at a high level of abstraction. This work targets models where parameters are only estimated with an interval (i.e. a lower and an upper bound). We first proposed a method to simulate exhaustively the state-space in the absence of feedback loop from the temperature model and the functional one, and explored possibilities to let the temperature model return concrete values to the functional model, and still provide a better state-space coverage than individual simulations.