Combining optimisation and simulation for steel production scheduling

Abstract

PurposeScheduling problems in steel plants tend to be difficult and require complex algorithms due to many constraints. An approach is presented where only the main constraints are included in the scheduling algorithm. The schedule is validated using a discrete‐event simulation model that includes additional detail.Design/methodology/approachThe combined approach is utilised for production scheduling in a steel mill in Finland. Operational performance of the steel mill is measured before and after software installation. The paper presents the scheduling environment, the software application and the resulting increase of production.FindingsCase experiences indicate that combining optimisation techniques with simulation is beneficial. The optimisation can be kept simpler as validation with a simulation model increases the credibility and accuracy of the resulting schedule. During software development and testing, the simulation model offered a testing environment for the optimisation algorithm.Practical implicationsThe case implementation was a success that increased production without making trade‐offs with other production goals. Company management estimate the productivity increase directly caused by the project to be worth €2,500,000 annually.Originality/valueThe paper presents a successful application of simulation for schedule validation in a complex and demanding environment.