This paper solves a scheduling problem of orders in a workshop with parallel machines, inhibition constraints in
manufacturing and setup times, with the objective of minimizing the mean tardiness in delivering those orders. To
solve the problem we applied a GRASP. In order to verify the proper use and efficiency of the proposed
procedure a computational experience was carried out on three data sets, with different number of machines and
kinds of jobs (or items) to be produced: 8 items and 3 machines, 12 items and 6 machines and 15 items and 9
machines. Each set has 100 instances, containing from 15 to 25 different orders, with their corresponding data.
The results were compared with those obtained using Genetic Algorithms. The GRASP has, in general, a better
performance than the Genetic Algorithm for the mean tardiness in the smaller instances. Nevertheless, for bigger
instances the Genetic Algorithm is more appropriated.
|