An Integrated Model for Simulating Activities and Optimizing Space Function in Educational Facility

Ren‐Jye Dzeng, Chin‐Teng Lin, Fan-Yi HsiaoInternational Journal of Information and Education Technology 2013 · DOI

Abstract

Determining the function for each space in a building plays an important role in assessing the service level of an educational facility. The function layout affects how users move between spaces during their participation in activities. This research describes the integration of an activity-based simulation model and a space-function assignment optimization model across six phases: data collection, data preparation, data mining, optimization, verification, and refinement. Using a real-world case study, the proposed model demonstrated a 4.96% improvement in objective value compared to a professional architect's original plan.

Introduction: The Challenge of Dynamic User Behavior

In traditional architectural design, focus often rests on static physical parameters such as lighting, temperature, and energy efficiency. However, the "service level"—how effectively a building supports the daily activities of its occupants—is significantly harder to quantify.

User behavior is inherently dynamic. Occupants do not remain stationary; they move between classrooms, laboratories, and offices based on complex, often overlapping schedules. Conventional design methods frequently overlook the "movement relation" between functions. For instance, placing a lecture hall far from its associated laboratory can lead to increased congestion and wasted travel time. This research addresses this gap by modeling user activities and movements to optimize space assignments during the design phase.

Methodology: The Integrated Model

The researchers developed a comprehensive six-phase framework to bridge the gap between activity simulation and spatial optimization.

The six-phase Integrated Model framework.
  1. Data Collection: Gathering information on "fixed" activities (scheduled classes) and "unfixed" activities (ad-hoc meetings, library visits) for various user groups.
  2. Data Preparation: Utilizing the Albatross model to generate realistic individual timetables based on probabilistic distributions.
  3. Data Mining: Analyzing these timetables to determine the "movement relation" ($R_{f_i f_j}$) between different functions, identifying high-traffic paths.
  4. Optimization: Implementing a Fast Messy Genetic Algorithm (fmGA) to assign functions to physical spaces. The objective function balances room-size suitability with the minimization of total movement distance.
  5. Verification: Using Cellular Automata (CA) simulations to visualize occupant trajectories in 2D and 3D, identifying potential corridor bottlenecks.
  6. Refinement: A final "human-in-the-loop" stage where designers adjust layouts for qualitative factors like noise interference or scenic views.

Case Study: National Chiao-Tung University

The model was validated using the 4-story Civil Engineering Department building at National Chiao-Tung University in Taiwan. The scope included 32 key spaces (classrooms, labs, offices, and meeting rooms) categorized by size. Data was derived from 510 students, divided into 58 distinct groups based on their specific course requirements and behavioral patterns.

Results and Discussion

The optimization process successfully navigated a massive search space of $2.631 \times 10^{35}$ possible combinations, converging on an optimal solution in just 250 seconds.

Optimized function assignments and movement relations.

Key findings include:

  • Efficiency Gains: The optimized layout (Result $R_1$) achieved a 4.96% improvement in the objective value over the original plan created by the contracted architect.
  • Functional Grouping: The model naturally grouped highly correlated functions (e.g., administration and the library) on the same floor and matched function requirements to appropriate room sizes.
  • Quantitative Verification: Cellular Automata simulations provided empirical evidence of improved service levels, such as reduced flow density in corridors and optimized occupancy rates.

Conclusion

While the integrated model offers a significant step forward in evidence-based design, some challenges remain. Qualitative constraints, such as the acoustic separation of offices from noisy classrooms, are difficult to quantify and currently require manual refinement. Future work may involve integrating real-time tracking data (e.g., Wi-Fi positioning) to further enhance the accuracy of activity models in even more complex, multi-use facilities.

Summary updated

Abstract

Determining the function for each space in a building plays an important role for assessing the service level of an educational facility. The function layout affects how users move between spaces during their participation of activities in the facility. This research describes the integration of an activity-based simulation model and a space-function assignment optimization model with six phases, including data collection, data preparation, data mining, optimization, verification, and refinement. The paper also uses a real case to demonstrate the use of the proposed model, and has shown a 4.96% improvement on the objective value compared to the contracted architect's original plan.
BibTeX
@article{dzeng2013integrated,
  author = {Ren‐Jye Dzeng and Chin‐Teng Lin and Fan-Yi Hsiao},
  title = {{An Integrated Model for Simulating Activities and Optimizing Space Function in Educational Facility}},
  journal = {International Journal of Information and Education Technology},
  year = {2013},
  doi = {10.7763/ijiet.2013.v3.340},
}
Figure