Wasp colony with a multiobjective local optimizer for dynamic task planning in a production plant

Abstract

Dynamic task scheduling is a time-dependent optimization issue. In this work, we modeled the process that is performed at a production plant as a task scheduling issue, in which a production line sends trucks to a painting plant with several stations. The objective is to attain efficient task scheduling, taking into account three conflicting objectives: number of color changes in booths, work tardiness, and makespan. In order to solve this problem, we developed a hybrid technique, which comprises a Wasp Colony algorithm and a set of priority rules. Both the problem and its solution were modeled through Agent Unified Modeling Languague (AUML) so as to achieve implementation. The results were a remarkable decrease in the number of color changes and work tardiness and the preservation of the number of painted trucks within an acceptable magnitude. © Springer-Verlag Berlin Heidelberg 2013.