Digital print workflow optimization under due-dates, opportunity cost and resource constraints

On-demand digital printing is an example of emerging personalized manufacturing services. It provides unique opportunities to automate the printing process, enhance productivity, and better utilize resources such as equipment, servers and IT infrastructure. In this work, we present a unified solution approach to solve an important optimization problem in digital printing, viz., simultaneous mapping of component tasks of a print job to time steps (scheduling), selection of resources for these tasks, and mapping of tasks to resources (binding). We model print jobs, the relationships between them, and dependencies between tasks within a job, in terms of sequencing graphs. This formal representation is then used for scheduling and resource binding. The optimization objective is to enable justin-time manufacturing, that is, to minimize both the slack time (the duration between the delivery deadline and the completion time of the order) and the opportunity cost for job orders. The proposed approach uses genetic algorithms (GA) to systematically search the space of feasible solutions. The fitness function of the GA is carefully crafted to match the optimization objective. An integer linear programming (ILP) model is described to evaluate the GA heuristic by deriving optimal solutions for small problem instances. The optimization technique is further evaluated using print orders from a commercial print service provider and compared to baseline methods commonly implemented in the industrial settings.