Robust product sequencing through evolutionary multi-objective optimisation

This paper describes a study on efficient optimisation of real-world product sequencing problems with the aim of finding robust solutions. Robust solutions are insensitive to unforeseen disturbances in a manufacturing process, which is a critical characteristic for a successful realisation of optimisation results in manufacturing. In the paper, the traditional method of achieving robust solutions is extended by using standard deviation as an additional optimisation objective. This transforms the original single-objective optimisation problem into a multi-objective problem. Using standard deviation as an additional objective focuses the optimisation on solutions that have both high performance and a high degree of robustness (that is, a low standard deviation). In order to optimise the two objectives simultaneously, a multi-objective evolutionary algorithm based on the Pareto approach is used. The multi-objective method for increased robustness is evaluated using both a benchmark problem and a real-world test case. The real-world test case is from GKN Aerospace in Sweden which manufactures components for aircraft engines and aero-derivative gas turbines. Results from the evaluation show that the method successfully increases the robustness while maintaining high performance of the optimisation. [Received 25 September 2015; Revised 18 October 2015; Accepted 5 November 2015]