Genetic algorithms for global sourcing

Abstract

Genetic algorithms have been used to solve a global sourcing problem, that is the allocation of production lots to different plants distributed all over the world in order to minimise production costs. Indeed some types of cars, like Fiat Palio and Siena, are called "world cars", as they are assembled in different sites in the world, often very far from each other and, in general, the components needed to assemble these cars are the same in all the locations. The manufacturer and its suppliers have to choose in which place to produce the components in order to optimise the global cost. Results obtained using genetic algorithms have been compared with a more classical technique based on a piece-wise linearisation of the global cost function. Results are reported related to some typical global sourcing problems.