Pervasive model adaptation: the integration of planning and information gathering in dynamic production systems

Abstract

Model-based planning often presumes a static system model, while in practice physical systems may evolve or drift over time. This paper proposes the idea of pervasive learning in a production system, where the model is dynamically updated using production output. The core idea is the interplay between model learning and production planning. We seek plans which simultaneously serve the goals of achieving high productivity for production and information gathering for model learning. The paper uses a simple modular printing example to illustrate issues such as formulation of the information criterion and search strategy for informative plans. The idea of pervasive learning can be further extended to achieve long term productivity in production systems.