Integrating agent-based modeling and dynamic lca: an environmental impact simulation and optimization model for dormitory buildings

Buildings are more than just shells of concrete and glass; they are living systems where the weather outside, the machines inside, and the people moving through the halls constantly influence one another. For decades, experts have tried to measure the environmental cost of these structures using a method called life cycle assessment. Think of this as a comprehensive audit that tracks every ounce of pollution a building creates, from the moment its materials are mined to the day it is demolished. Traditionally, these audits have been static snapshots, taking a single average year and calculating the impact based on fixed assumptions. However, the real world is rarely static. The temperature shifts hour by hour, appliances get older and less efficient, and the people living inside change their habits depending on the day of the week or their mood. When researchers ignore these shifting dynamics and the way they interact with each other, their calculations can drift far from reality, leading to management strategies that miss the mark.

To bridge this gap, a team of researchers from Southeast University, the University of Wisconsin–Madison, and Dalian University of Technology has developed a new way to simulate the life of a building. They combined two powerful tools: a method that tracks how individual people and objects behave over time, and the environmental audit mentioned above. By merging these approaches, they created a model that does not just look at a building in isolation but watches how the climate, the building's equipment, and the students inside it react to one another in real time. They tested this system on graduate dormitory buildings in Nanjing, China, a setting chosen because these high-density living spaces consume significant energy and offer a clear view of how human behavior drives environmental impact.

The researchers built a digital twin of the dormitory system where three types of virtual agents interact. First, there is the climate agent, which feeds in hourly temperature data to see how the weather pushes students to turn on heaters or air conditioners. Second, there is the dormitory agent, representing the physical rooms and their aging appliances, which become less efficient and use more power as they get older. Third, and perhaps most importantly, is the student agent. These virtual students have schedules, habits, and a level of "green consciousness" that dictates how carefully they use electricity. The model simulates how a student might decide to turn off a light, how a group of roommates might influence each other's energy use, and how these small decisions ripple out to change the building's total environmental footprint. Crucially, the model includes a feedback loop. It does not just calculate the damage; it allows a virtual administrator to try out management strategies, see the results, and adjust the plan accordingly, creating a continuous cycle of assessment and improvement.

When the team ran their simulations, they discovered that treating dynamic factors as separate, independent variables leads to significant errors. If you look only at how temperature changes, or only at how students behave, without seeing how they affect each other, your assessment can be off by as much as 226,000 units of environmental impact. The model revealed that the highest environmental impacts did not always happen when the weather was hottest or coldest. Instead, peaks occurred when extreme weather coincided with high occupancy, such as during the start of the school year, while impacts dropped sharply during holidays when students left campus, even if the weather remained extreme. This confirmed that the interaction between the people and the environment is the true driver of energy use, not just the weather alone.

The study then tested two specific strategies to see if they could lower these impacts. The first was a campaign to improve the students' green consciousness through education and competitions. The second was a logistical change: reorganizing the dormitory assignments so that students of the same academic year lived together, rather than mixing freshmen and seniors in the same rooms. The results showed that both strategies worked, but in different ways. Improving green consciousness reduced the total environmental impact by about 2.7 percent, creating a steady, uniform improvement across all buildings. The dormitory reorganization was even more effective, cutting the cumulative impact by over 5 percent. This was because students in the same grade tended to have similar schedules, which meant they turned lights and computers on and off at the same times, reducing the hours that shared equipment had to run unnecessarily.

Perhaps the most revealing finding was that the starting point of the system matters immensely. The researchers found that if students already have high green awareness, the room reorganization strategy still provides a significant boost, proving that good management is always necessary regardless of how "green" the occupants are. Conversely, if the starting conditions are poor, the potential for improvement is larger, but the final result depends on how quickly those conditions can be changed. The study also highlighted that ignoring the aging of equipment leads to an underestimation of future energy use, as older machines require more power to do the same job. By integrating these factors into a single, looping simulation, the researchers demonstrated that environmental management is not a one-time calculation but a dynamic process. Their work suggests that to truly reduce the environmental burden of buildings, we must stop looking at weather, machines, and people as separate lists and start understanding them as a single, interacting system that can be guided toward a better future.