Including real-life problem preferences in genetic algorithms to improve optimisation of production schedules

This paper looks at extensions that can be made to genetic algorithms (GAs) used for schedule optimisation when meeting the additional complexities provided by working with industrial applications. The optimisation process should take into account the preferences of the users involved in implementing the scheduling process. Problems are provided by the challenging real-life manufacturing problem based on a chilled ready meal factory. The additional amount of data found in a real-life problem can be combined with the power of the GA optimisation process to improve the quality of solutions rather than limit the potential for the user. The inclusion of factory preference data in a schedule builder, and the interaction with a multiobjective GA system allows the users´ preferences to be assembled into successive stages of the optimisation process to meet this requirement. By extending standard GA implementations to adopt the inclusion of user preferences, schedule optimisation system can demonstrate the full capabilities of GAs for tackling difficult industrial problems such as manufacturing schedule optimisation