Combining heristic and utility function for fair train crew rostering

• Autor(a) principal: Constantino, Ademir Aparecido
• Data de Publicação: 2015
• Outros Autores: De Mendonça, Candido Ferreira Xavier, Novaes, Antonio Galvão, Dos Santos, Allainclair Flausino
• Tipo de documento: Artigo de conferência
• Idioma: eng
• Título da fonte: Repositório Institucional da UNESP
• Texto Completo: http://hdl.handle.net/11449/220432
• topic: Bottleneck assignment problem; Crew rostering problem; Fair rostering; Stated preferences

In this paper we address the problem of defining a work assignment for train drivers within a monthly planning horizon with even distribution of satisfaction based on a real-would problem. We propose an utility function, in order to measure the individual satisfaction, and a heuristic approach to construct and assign the rosters. In the first phase we apply stated preference methods to devise a utility function. The second phase we apply a heuristic algorithm which constructs and assigns the rosters based on the previous utility function. The heuristic algorithm constructs a cyclic roster in order to find out a minimum number of train drivers required for the job. The cyclic roster generated is divided into different truncated rosters and assigned to each driver in such way the satisfactions should be evenly distributed among all drivers as much as possible. Computational tests are carried out using real data instance of a Brazilian railway company. Our experiments indicated that the proposed method is feasible to reusing the discrepancies between the individual rosters.