System-Theoretic Properties of Production Lines.

Abstract

Production lines with unreliable machines and finite buffers have been studied from the point of view of performance analysis and optimization for over 50 years. In contrast, system-theoretic properties of production lines have rarely been analyzed. These properties, however, are of significant practical and theoretical importance because they reveal the fundamental principles that characterize the behavior of such systems and can be used for their management and improvement. This dissertation is devoted to the investigation of system-theoretic properties of production lines in large volume manufacturing industries. In Control Theory, the main system-theoretic properties addressed are stability, controllability, and observability. For production systems with Markovian dynamics, the stability is ensured by the properties of the eigenvalues of the state transition matrices. As for the controllability and observability, since active feedback control is practically never used to manage production lines, they are not of immediate industrial importance. Therefore, we focus on system-theoretic properties, which are central for the management of production lines. They are: reversibility, monotonicity, improvability, effects of up- and downtime, bottlenecks, and transient behavior. For each of them, we derive tests for existence and demonstrate their applicability in case studies at automotive plants.