Neuro-tabu search approach to scheduling in automotive manufacturing

Abstract

Abstract The work concerns a new problem of minimizing production time of a set of tasks on machines while limiting the availability of machine operators. The problem is taken from practice, from a company producing subassemblies for the automotive industry. A mathematical model of the problem is formulated, and for its solution a neuro-tabu algorithm is proposed, which is a variation of the tabu search algorithm with a neural mechanism replacing the tabu list: short-term memory. The e_ciency of the algorithm constructed in this way is veri_ed on test examples based on real data from the production process.