Optimization of Finite Capacity Scheduling based on an Evolutionary Approach

H.O. Nyongesa (South Africa), S. Maleki-dizaji (UK), and Z.A. Mbero (Botswana)


Production planning, finite capacity scheduling.


In this paper we discuss an optimization approach for a real-world production planning and scheduling problem. Based on data from real instances of production planning an algorithm and architecture were developed for optimization of production planning and scheduling for manufacturing lines in small and medium enterprises (SME). The optimization approach is based on the use of genetic algorithm.(GA). The approach referred to as Unified Planning using Intelligent Techniques, UPlanIT is based on genetic algorithms (GA), which are evolutionary machine learning techniques. The optimized schedules are constructed using rules in which priorities are determined by the GA, using a procedure that generates parameter constrained activities. Furthermore, after a schedule is obtained local search heuristics can be applied to improve the solution, either automatically or manually. The approach is tested on a set of standard production instances and is compared against a legacy optimization system in operation at a case study SME. The results show the evolutionary approach to achieve a higher efficacy in optimization of production planning.

Important Links:

Go Back