The NSF Sustainable Energy pathways (SEP) Program, under the umbrella of the NSF Science, Engineering and Education for Sustainability (SEES) initiative, will support the research program of Profs. Burcin Becerik-Gerber of the Department of Civil and Environmental Engineering, Wendy Wood of the Department of Psychology, David Gerber of the Department of Architecture, and Milind Tambe of the Department of Computer Science at the University of Southern California (USC). The multi-disciplinary team of investigators will develop an energy-aware, cyber-physical multi agent framework of buildings, humans, and intelligent software agents for sustainable energy management, taking a collective, energy literacy approach to influencing building occupants, operators, designers, and engineers. The investigators will first assess behavior and preferences of building occupants, evaluate building design/system specifications, and identify building operational policies. They will then build a multi-agent model to integrate these different systems. Building on fundamental research in agents' autonomy and teamwork, the multi-agent framework will facilitate negotiations between occupants and building devices. The agents will provide feedback to the occupants and control building devices to conserve energy. Based on this integrated model, feedback about occupant energy use to building designers will be provided to shape early-stage design decisions that have the longest lasting impact on building's lifecycle footprint. The central focus is designing a multi-component model of energy consumption in office buildings in order to identify and test the optimal points of change in energy systems. Specifically, the research predicts that energy use could be optimized and occupant comfort could be maximized in an integrated way by changing occupant behavior, design/system specifications, and building operators' policies via an agent-based system. The research will be validated in an office building, where occupants lack the individual financial incentives for energy consumption. The system will be tested both in professional and student designer studios to validate the impact of the model in energy aware design decisions. The research differentiates itself by treating occupant preferences and behavior not only as input data but also as controllable variables in a broader energy system; it then harnesses a complex multiagent system to control these variables for energy savings. It also extends energy literacy into the arena of design and engineering by providing human behavior input in early design stages, as well as into the arena of building operations by dynamically controlling buildings based on human behavior and preferences.
With respect to the increasing energy needs of our country and world, this research has far-reaching impact on environmental conservation, pollution, and the economy. The primary impact of this research is identifying the key factors that create significant energy savings for buildings, resulting in monetary savings and environmental protection. The results will be disseminated through and contribute to multiple conference talks and publications. A game, in which students will compete to save the most energy, will be developed with the aim of teaching how to conserve energy in daily life. Energy-focused workshop lessons will be developed and delivered to minority-majority K-12 schools in USC's neighborhood and other children and their families. In addition, the research team will partner in research with the computer science department of California State University, Dominguez Hills, a minority majority institution.
The proposed interdisciplinary research challenges the ways that building engineers and designers, computer scientists, and behavioral scientists look at the pressing challenges of energy-efficient buildings. The research will integrate design with outside data sources and bring behavioral science to design, which in turn will trigger a design-method evolution for sustainable buildings. Through dynamic data collection, spatiotemporal information about energy use, and data on human comfort, the research will improve system optimization and adaptation through the use of intelligent software agents. The project will educate building occupants about their energy consumption and the ways that they can make concrete behavioral changes to achieve greater energy efficiency. The research brings to the problem of energy literacy an interdisciplinary approach in which cyber tools are manifested in physical space.
