Iterated local search for the label printing problem

In the label printing problem (LPP), a set of labels must be printed in specified quantities, using n templates to be defined. Each template can accommodate a fixed number of printing plates. The LPP consists of determining: (1) an n-partition of the labels, where each block defines the labels that share a template, (2) for each label the number of identical printing plates to place in one of the n templates, and (3) the number of imprints to make with each template. The objective is to satisfy the labels' demand while minimising total waste. Since the LPP is NP-Hard, to solve it, we propose here an Iterated Local Search heuristic that, once tested on all the instances available to us, it improved many of the best known results or find the optimum, when it is known.