Libraries are hearts that pump knowledge into the university community, and as such are important pillars of student life. The Linderman and FML libraries provide an alluring environment to the university's diverse student body for research, study and collaboration. Managing such facilities is an endeavoring task and hence libraries often seek help from the student community. Around 40 students are offered on-campus employment at circulation desks to cover about 112 shifts throughout the week at both the libraries.
We both worked as circulation supervisors last fall and had realized the difficulty of the shift scheduling problem at hand. During the Fall 2019 semester, we took the course ISE 426: Optimization Models and Application instructed by Professor Albert S. Berahas. As part of the coursework, we were tasked with identifying a real-world problem, formulating a mathematical model and solving it using the optimization techniques discussed in the class. We instantly knew our problem statement: Library Shift Scheduling. Shift scheduling is a combinatorial problem with many intricacies. Our solution, optimization, and specifically, integer programming.
We had numerous discussions with Professor Berahas, who guided us in developing a systematic approach towards the problem. After successful implementation on simulated data, we approached Kathleen Dugan, the Linderman Library coordinator, to prepare a shift schedule for the Spring 2020 semester. Data collection was a colossal task; our model and solutions were only as good as the data at hand and we had to preprocess the information collected. The problem was solved using AMPL, a popular optimization tool to create shift schedules which were implemented during the Spring semester.
Kathleen Dugan, stresses: “Usually a lot of iterations are required to create the final semester schedule, but this time the students too were satisfied by the initial schedule created using the model”. In addition to the optimal schedule, we provided a backup schedule, and a general availability schedule to address the issues of unattended shifts and replacements. The 3-schedule system turned out to be very effective for the coordinator when assigning shifts. “This is the first time in 20 years that I didn’t have to create a schedule by myself”, said Kathleen Dungan who was excited about the prospect of adopting this scheduling system as a full-time solution.
We look forward to extending the model’s application for other Lehigh University facilities. We are currently working on developing an end-to-end software interface which would enhance the user experience and streamline the entire process. We had the opportunity to present (see poster and slides) our work at the Virtual Lehigh Expo. We are grateful to have been awarded the first prize under the Master’s category for presenting our work at the ISE Undergraduate and Master’s Research Symposium 2020.
Overall, the process of solving the Lehigh University’s Library Shift Scheduling problem helped us realize the power of methodological and systematic approach to solving real-world problems and the importance of mathematical foundations and tools. We are very grateful to the ISE department, and Lehigh University, for helping us hone our analytical and problem solving skills and look forward to our next adventures at Lehigh University and beyond.
