Improved Staff Rostering Algorithms
Isaac D. Cleland, Andrew Mason, and Michael O’Sullivan
Department of Engineering Science, University of Auckland
The Staff Rostering Problem involves optimising the assignment of staff to shifts, while fulfilling various rules and preferences associated with these assignments. These problems are difficult to solve automatically; they are NP-Hard, regularly have a large number of variables, have a non-linear or multi-objective cost function, and are tightly constrained. We have developed a suite of algorithms to solve these problems using both Column Generation (Branch and Price) and Column Generation based Matheuristics which we will outline in this presentation.
The International Nurse Rostering Competition (INRC) was a global competition held in 2010 to benchmark the best Staff Rostering solvers against each other on a large number of difficult- to-solve Nurse Rostering Problems. We will present some preliminary results from our suite of algorithms in solving the INRC problems together with some real world case studies.
This presentation is eligible for the ORSNZ Young Practitioners Prize.