Comparison of different evolutionary algorithms for global supply chain optimisation and parameter analysis

Abstract

This paper discusses global optimisation from a business perspective in the context of the supply chain operations. A two-silo supply chain was built for experimentation and three approaches were used for global optimisation: a classical evolutionary approach, a cooperative coevolutionary approach and a coevolutionary approach with on the fly partner generation where the solution from the second component of the supply chain is generated deterministically based on the first one. The second approach produced higher quality solutions due to its use of communication between silos. Additional experiment was conducted to choose optimal species sizes.