Nurse Preference Rostering using Agents and Iterated Local Search

Document Type


Publication Date



This study presents an iterative local search (ILS) framework used within an agent-based nurse rostering system. This agent based system with the ILS (CNR-ILS) is used to improve nurse rosters with respect to nurse preferences. The system is a heuristic that considers a variety of nurse considerations including informal requested days off, preferences for days-of-the-week off, and preferences for specific numbers of consecutive days off. Since our implementation is not subject to the complexity of mathematical programs, it includes more preference considerations. The system is extensively tested in a federal hospital and is compared to both an integer programming solution from the literature and the scheduling methodology currently used at the test facility. When nurses are surveyed, the solutions found by this system are consistently favored over the solutions from the integer program and hospital’s current scheduling practice. Furthermore the variation of those ratings is significantly less.


The "Link to Full Text" button on this page loads a view-only version of the article by Springer Nature SharedIt.

AFIT users may access the institutional subscription article full-text by clicking here.



Source Publication

Annals of Operations Research