|
Abstracts
Contributed Talks
On the Multi-objective Nature of Timetabling and Rostering Problems
Dario Landa Silva
Timetabling and rostering problems are frequently formulated as multi-criteria decision
problems. We propose a procedure to assess the multi-objective nature of these problems and
illustrate it using a number of test cases. This allows a better guided approach for tackling
these problems in a multi-objective fashion.
A Model for Distributed Employee Scheduling
Patrick De Causmaecker
In distributed employee scheduling autonomous departments manage detailed schedules for their
assigned employees while the company as a whole will profit from a more globally optimised
schedule. Departments do not normally need to communicate in detail but can profit from
exchanging aggregated information. A model and experimental results are discussed.
Comparison of Solutions from a Genetic Heuristic and Goal Programming for Nurse Rerostering with Soft Constraints
Margarida Pato, Margarida Moz
The nurse rerostering problem was modelled as an integer multicommodity flow bi-objective
problem. A simple genetic heuristic and a goal programming approach are presented, along with
results of a computational experiment designed to compare these methods within a set of
instances taken from a real situation at a Lisbon hospital.
Adaptive Driver and Bus Scheduling
Vitali Gintner, Natalia Kliewer, Ingmar Steinzen, Leena Suhl
Traditionally, driver and bus scheduling are solved sequentially. Consequently, the generation
of feasible duties is restricted by vehicle blocks which results in infeasible or suboptimal
driver schedules. The proposed adaptive approach allows to recombine the given vehicle blocks
during driver scheduling resulting in still optimal vehicle schedules and improved driver
schedules.
New Evolutionary Algorithm for Examination Timetabling Problem
Zahra Naji Azimi, Majid Salari
In this approach we modify Scatter Search with our new methods and solve the Examination
Timetabling Problem. Also we solve this problem with existing methods such as SA and TS and
compare results of them with each other. Finally we Statistically conclude that our algorithm
works better.
An Investigation on High Level Heuristics within a Graph Based Hyper-heuristic for Course and Exam Timetabling Problems
Rong Qu, Edmund Burke
This paper presents our work on investigating different searching methods within a graph based
hyper-heuristic. Fundamental issues concerning the neighborhood structures are discussed and
performance within the landscape of search is analysed. Experimental results on both course and
exam timetabling problems demonstrate the simplicity and efficiency of this hyper-heuristic
approach.
|
