A Mixed-Integer Optimization Tool for System-Level Proactive Obsolescence Risk Management

Abstract

Diminishing manufacturing sources and material shortages (DMSMS) pose a significant risk to the reliability and maintainability of many modern complex systems due to part obsolescence. Part obsolescence increases the risk of system failure or downtime, which is often not only undesirable, but non-negotiable. As such, it is imperative to anticipate events of part obsolescence and take proper actions to mitigate their effects. However, developing a detailed system-wide plan of mitigating events of obsolescence could necessitate the consideration of thousands, or even millions, of decisions before an optimal plan is derived. Thus, there exists a need for an automated tool that can produce a cost optimal obsolescence management plan. In this work, we propose a mixed-integer optimization tool that proactively manages and mitigates events of obsolescence at the system level. By considering the architecture of the bill of materials, part inventories, expected system demand, and forecasted obsolescence dates, our tool makes decisions of when and how to enact up to 8 unique short-term and long-term resolutions for every part within the system at every time until the anticipated system end of life. The parameters determine the structure of a mixed-integer optimization problem, where the objective is to minimize costs for the system over the entire time horizon. We present the findings of a case study that compares the obsolescence management plan of a reactive strategy to a plan produced by our mixed-integer programming optimization tool and derive insights based on the difference between the solutions produced by the two methods.