The Department of Industrial Engineering is pleased to welcome Dr. Kelly Sullivan from the University of Arkansas’ Industrial Engineering Department. Dr. Sullivan will be the special guest lecturer on Monday, October 20, 2014 from 11:15am until 12:05pm at the P&A Auditorium on Clemson’s main campus. The topic of is lecture is: Integer-Linear Models for Optimization in Reliability.
Guest Lecturer: Dr. Kelly Sullivan, University of Arkansas, Industrial Engineering Department
Date: Monday, October 20 at 11:15am to 12:05pm
Location: Poole Hall, A102 Poole Hall (P&A Auditorium)
Title: Integer-Linear Models for Optimization in Reliability
Abstract: The problem of allocating limited resources to design, configure, and maintain a system for maximum reliability has important applications in a variety of industries. We present a general, exact optimization technique for this class of problems based on representing system reliability through a linear program (LP) that arises when the system is represented as a binary decision diagram (BDD). This LP extends intuitively to formulate the design, configuration, or maintenance problem as a mixed integer linear program (MILP). Computational results demonstrate that the BDD-based MILP can be solved efficiently for structured (e.g., series-parallel) systems.
We then demonstrate how this approach can be used to allocate limited testing resources in the development of a product (e.g., defense systems). Testing such a product identifies possible causes of failure whose removal results in improvement of system reliability, or “reliability growth.” We model the problem of allocating testing resources to maximize reliability growth, and we develop and analyze an exact solution approach for this problem based on solving a sequence of BDD-based MILPs.
Bio sketch: Dr. Kelly Sullivan is an Assistant Professor in the Industrial Engineering Department at the University of Arkansas. His methodological areas of expertise include integer programming and network optimization. His research interests, including network interdiction and optimization in reliability, center around developing and applying optimization methodology in the design of resilient systems. Dr. Sullivan joined the faculty at the University of Arkansas after completing his Ph.D. at the University of Florida in 2012.
