Seznam |

Téma: | Variable depth search for nurse rostering |
---|---|

Vedoucí: | Přemysl Šůcha |

Vypsáno jako: | Práce v týmu a její organizace |

Popis: | The Nurse rostering Problem (NRP) is a well-known NP-hard combinatorial problem dealing with the management of human resources in a hospital. The goal is to provide as good final rosters as possible while many constraints have to be considered. Good rosters have a significant impact on the quality of health care, the satisfaction of nurses, the hospital budget, etc. This problem can be solved by many different algorithms, both exact and heuristic. The exact algorithms produce optimal solutions but can take really long time to finish. On the other hand, the heuristic algorithms return sub-optimal solutions within the reasonable time. Therefore, the main aim of this project is to implement a heuristic algorithm (and all necessary parts around) based on [2]. |

Pokyny: | The project contains the following tasks:
1. Becoming acquainted with the nurse rostering problem. 2. Design of data structures and implementation of a parser for the input instances in the format described in [1]. 3. Design and implementation of the procedure, which evaluates the intermediate solutions, based on the constraints obtained from the previous point. 4. Implementation of the variable depth search algorithm [2] with the specific modifications from [3]. 5. Verifying the proper functionality on the standard benchmark instances. 6. Displaying the detailed output (e.g. HTML), i.e. the resultant roster together with the marked constraints which were violated (including the description why the particular constraint was violated). The project should be implemented in Java/C#. |

Literatura: | [1] T. Curtois: Staff rostering benchmark data sets. http://www.cs.nott.ac.uk/~tec/NRP/, 2015. [2] E. K. Burke, T. Curtois, R. Qu, G. V. Berghe: A Time Predefined Variable Depth Search for Nurse Rostering. INFORMS Journal on Computing, 411-419, 2013. [3] E. K. Burke, T. Curtois: New approaches to nurse rostering benchmark instances. European Journal of Operational Research, 71-81, 2014. |

Realizace: | SW project |

Vypsáno dne: | 10.02.2015 |

Max. počet studentů: | 5 |

Přihlášení studenti: | |