A54 Postgraduate Statistics Centre
10am-4pm (Buffet lunch provided 12-1pm)
Experiment Design & Analysis for Dynamic, Stochastic Simulation
with Professor Barry L Nelson, Northwestern University, USA
Abstract: The analysis of systems via discrete-event, stochastic simulation involves modeling and a statistical experiment. This masterclass covers the design and analysis of that experiment, emphasizing aspects that are central to simulation experiments but less common in physical experiments. After a brief review of simulation and some mathematical/statistical basics, we address the key distinction between risk and error; the impact of using estimated input models to drive the simulation; control of error in simulation estimates; and simulation optimization (i.e. an optimization problem in which some simulation response is the objective to be maximized or minimized). The class assumes only a first course in simulation that emphasized modeling.
To download presentation slides for this masterclass please click on the links below: