Stochastic simulation is a powerful modeling and software tool for analyzing many modern complex systems in automation. Optimization involving millions or even billions of decision variables and constraints has become possible with the advent of modern computing. The combination of these two successful paradigms, called stochastic simulation optimization, has become dramatically powerful for addressing practical systems in recent years. However, computational efficiency is still a big concern because i) in the optimization process, many alternative designs need to be simulated; ii) to obtain a sound statistical estimate, a large number of simulation runs (replications) is required for each design alternative. A user may be forced to compromise on simulation accuracy, modeling accuracy, and the optimality of the selected design. There is a critical need for a technological breakthrough in simulation optimization methodologies.
Many interesting such examples can be found in control and optimization problems of smart buildings. Buildings are responsible for nearly 40% of the energy consumption in the United States as well as many other developed countries, and are responsible for nearly 30% of the energy consumption in China as well as many other developing countries. Comparing with industries and transportation systems, buildings have the largest energy saving potential. This may be achieved through better human-building interaction, better design of the various devices in buildings, and better joint scheduling of these the multiple energy systems in buildings. These problems usually involve simulation or physical experiments, where simulation optimization methodologies play important roles.
This session intends to provide some recent development in simulation optimization research in general, and to provide their applications in various interesting problems in smart buildings.