Model for Prioritization of High Variation Elements in Discrete Production Systems

The complexity of the modern manufacturing enterprise has led companies to look for techniques and methodologies for improving production performance. Lean manufacturing techniques have been applied in the US with varying degrees of success, and Theory of Constraints (TOC) has been used to emphasize the flow of production and identify performance improvement projects. One aspect of manufacturing for which there has been limited academic or industrial research till date is the impact of variation on production performance and the identification of improvement projects based on variation. This thesis develops a methodology to incorporate random and simultaneous occurrence of variability in a manufacturing facility, e.g., equipment failure, variabilities in the arrival time of raw materials and in-station processing time, to model system performance. Two measures of performance are developed corresponding to time and material. A prioritization algorithm is developed to utilize the “Coefficient of Variation” to identify a Bundle of High Variation Elements (BHVs) affecting the performance of a production system. The Bundled Variation-based Project Prioritization Model (BVPM) is a closed-loop model designed to provide decision makers with a list of projects to improve system performance while monitoring the implementation of projects.