Use of a genetic algorithm in a distributed resource management software

Resource management and planning in railway operating companies is a highly complex task. Based on simulation, service planning, vehicle rotation planning, personnel rotation planning and operations management, the resources (e.g. people, vehicles, and infrastructure) must be combined to timetables and rosters. In this work, the development of an optimization agent, using a genetic algorithm, for an existing distributed software product for resource planning (RailOpt) is treated. Main aspects are the description of an underlying real world problem, the definition of a fitness function in order to solve a sub-problem, the particular implementation of the GA and its configuration.