Poster
Client Context
Georgia Tech Capital Planning and Space Management (CPSM) helps provide space solutions to Georgia Tech’s 240 space users for the 3.3 million square feet of the 12,000 instructional, research, and office spaces on Georgia Tech’s campus. One of these space solutions is quantifying availability of swing spaces (defined in next section) when renovations cause a building to go offline, which then informs the timeline, budget, and which contractor will complete a renovation. This renovation plan must be approved by the Registrar, individual departments, and the Executive Leadership Team.
Project Objective
The planning of renovation budgets and timelines requires CPSM to work with the impacted departments and the Registrar to locate sufficient swing space, available space that can temporarily be reallocated to accommodate displaced courses during the duration of the renovation, to ensure that course loads can be maintained during the construction period. This search for instructional swing space determines project timelines and is typically the deciding factor for whether a renovation is completed in one continuous single phase, or if it must be divided into multiple phases to avoid taking all of a building’s instructional spaces offline at once. If feasible, single-phased construction is preferred due to convenience and significant cost savings.
The problem is that the Registrar must manually rearrange courses, or even give a ‘gut’ response, to determine if a single-phased renovation is feasible. CPSM must also manually search for swing space solutions by calling campus contacts through a process of gradual discovery.
Design Strategy
We worked with CPSM to understand the challenges and requirements in planning the renovation of the Howey Physics building. Since this building has four large capacity lecture halls, it served as a good case study in developing the design. Large capacity lecture halls are a limited resource, so planning the Howey renovation was more challenging than planning the renovation of other buildings. The Howey renovation was completed in three phases.
In our design, we focused specifically on displaced instructional space. The tool we built had to be flexible, and able to be used in an iterative process, to balance the preferences and needs of the stakeholders (Executive Leadership Team, Departments, and the Registrar). When relocating courses that are displaced, there are certain things that must be met, like course capacity being less than room capacity, wheelchair accessibility, and distance learning equipment being present if needed. Similarly, classes being displaced from lab spaces need to be kept in lab spaces. There are also some compromises that can be made, such as reducing course capacity, moving courses online, and moving classes to departmentally scheduled rooms.
Based on these needs, we built the Swing Space Analysis Tool, a constraint programming model with an objective of minimizing the number of classes moved online, to maximize the number of displaced courses that are assigned a physical room. The hard and soft real-world constraints are incorporated into the model, and the tool can be used iteratively. It uses historical course data to determine the feasibility of completing the renovation in that semester. If there is planned course offerings data for an upcoming semester, that can also be used as an input. With each output of the tool, CPSM can discuss with the affected stakeholders and adjust the inputs to the tool to run it again.
As we built the tool, we worked with CPSM to see how they would use the tool in planning the past Howey Physics building renovation and the upcoming D.M. Smith and Van Leer building renovations.
Deliverables
Swing Space Analysis Tool: To use the Tool, CPSM will collect historic semester schedule data from Banner. They can then adjust the flexible model constraints through input spreadsheets, to reflect the preferences and compromises of the various decision makers, such as the BME department preferring their courses to be offered in the BioTech Quad, or the Paper Building being unable to offer courses on Fridays. A Python script then reads these input spreadsheets and communicates the objective and constraints to a CPLEX solver to find a solution that maximizes the number of courses that are assigned to a physical swing space solution. Its secondary objective is to minimize the amount of reduced course enrollment capacity. The solver’s solution is outputted in an Excel spreadsheet that includes: The new room for each displaced course, or a remote delivery recommendation if a room cannot be located; and The reduced capacity of each displaced course, if necessary. This capacity reduction is subject to both a maximum drop percentage (in regard to percent of the enrollment capacity) and a maximum number of courses that can receive a capacity reduction
Tableau Dashboards: CPSM will upload the Excel spreadsheet produced by the Swing Space Analysis Tool into Tableau to populate the Displacement Summary Dashboard. This dashboard allows CPSM to quickly identify the courses that have received a capacity reduction, been moved to an alternate time, or been recommended for remote delivery. This list of courses can be filtered by department to measure inconveniences when considering individual department preferences. The dashboard can also be used to evaluate the additional student and course hour load taken on by the campus buildings that accept displaced courses.
Future Recommendations:
Currently, data that shows space ownership, including unassigned spaces, is tracked every two years. To improve upon the collection of data that would be used in the Swing Space Analysis Tool, our team recommends more frequent tracking of other spaces that could offer courses so that these rooms could be added to input data if an outputted solution recommends too many courses be offered online.
Secondly, the transition to mostly online classes with the on-going COVID-19 pandemic has made more data available for which courses have been most successfully transitioned to remote delivery. We recommend that CPSM compiles a list of these courses that have minimal academic when offered in a remote format. Therefore, if future renovation projects require moving courses online, these courses can be easily communicated to the model so that courses that are not as successful when offered in a remote format will be less likely to be recommended for remote delivery.
Project Information
Student Team
Prerna Balaji, Savannah Chunn, Zach Hess, Chidambaram Kadiresan, Makala Muhammed, Abbey Nannis, Sarah Poff, Sena Sennaroglu
Faculty Advisor
Faculty Evaluator
