(This presentation by Peter May-Ostendorp will begin with an introduction to building energy research at UC Boulder, by Prof. Gregor Henze.) Mixed-mode (MM) cooling is a promising building design strategy for low-energy cooling that incorporates natural ventilation alongside other forms of space conditioning. A properly designed system will intelligently switch between modes of cooling to maximize energy savings, while preserving occupant comfort. The near-optimal operation of MM buildings is explored through a model-predictive control (MPC) study using a purpose-built optimization environment coupled to EnergyPlus. Preliminary results indicate that MPC is effective at detecting viable control schemes for MM building operation, such as night cooling, and may provide a reasonable benchmark against which to critique the optimality of simplified, heuristic control designs. Since MM building performance is heavily impacted by occupant behavior, methods for modeling simplified occupant behavior through existing EnergyPlus routines as well as an outlook on probabilistic occupant modeling are presented. A general linear model (GLM) method for extracting control rules from the near-optimal results is in its early stages, and this technique may provide the ability to control MM buildings through simplified but, nonetheless, near-optimal control rules without the need for computationally burdensome, real-time MPC.