Multiobjective Optimization of the Train Staff Planning Problem Using NSGA-II

Abstract

The optimization problem of assigning train staff to scheduled train services is called the train staff planning problem. A part of this is the rostering with the aim to create a duty timetable under the consideration of different constraints, preferences etc. The problem is formulated as a biobjective problem considering costs and penalties for violating constraints. In this paper, we analyze the application of the nondominated sorting genetic algorithm II (NSGA-II) for multiobjective optimization in order to propose a solution to the considered train staff planning problem. Numerical experiments are conducted using several example problems. These experiments provide suitable parameters for using NSGA-II and further insights into the adaptation of this algorithm to the problem under consideration.