Temple Dental School Uses an Expert System to Schedule Students' Clinical Rotations

Each semester, Temple University School of Dentistry must develop a complex schedule for students' clinical rotations in seven clinics. The schedule must satisfy several demands: clinics have to be adequately staffed, students need time off for lectures and treating patients, and each student should serve approximately the same number of rotations over a two-year period. A number of methods are available to solve scheduling problems such as this. They range from simple rules and linear programming to custom-coded algorithms and expert systems. The many, sometimes conflicting requirements of the schedules at Temple Dental School made using mathematical or algorithmic methods difficult. I developed a scheduler in a rule-based artificial intelligence language (Prolog) to generate schedules that satisfy all user requirements. It has resulted in yearly savings of 11 to 19 person-days for programming and schedule development.