Constraint programming and the physician rostering problem

Abstract

The Physician Rostering Problem (PRP) seeks to assign shifts to physicians so that all the demands of the hospital are met, no physician is overloaded, and the scheduling is as pleasant as possible for them. Constraint Programming (CP) is a paradigm for solving combinatorial problems, which combines techniques of Artificial Intelligence, Theory of Computation, and Operations Research. This project aims to study CP and its use in the design of algorithms for PRP and related problems, as well as the implementation and testing of some of these algorithms. In PRP there are rigid restrictions, which must be met, and flexible restrictions, which improve the quality of the solutions when they are met. We intend to approach the quality and variety of solutions from the perspective of multi-objective optimization. The project also aims to complement the candidate's training in the areas of combinatorial optimization and algorithm design and analysis, in addition to introducing him to research work.(AU)