Abstract: Correlation in disruption risk is a critical factor to be considered when designing systems like warehouse locations, inventory policies, etc. However, modelling correlation in risk is a problem fraught with the curse of dimensionality where the number of parameters increases exponentially with the number of dimensions of risk one needs to consider. In this talk, we introduce a novel methodology to model correlation in multi-dimensional risk. Our approach, called subordinated Markov chains, is parsimonious, effective and amenable to optimization thus making it an apt choice to be deployed when considering decision-making in an operations management context, in the presence of correlated risk. We demonstrate the effectiveness of the methodology through the facility location problem with correlated disruptions. We also extend the model to inter-temporal correlation of disruptions and show its efficacy in modelling delay propagation in airline networks.

About the Speaker: Prof. Vishwakant Malladi obtained his PhD in Risk and Operations Management from the McCombs School of Business, UT Austin. His research primarily focuses on risk in an operations management context and can be broadly divided into two areas. First, he works on parsimonious modelling of risk in high dimensional systems using Levy processes. Second, he studies the impact of risk and risk correlation in operations management problems like inventory theory, reliability and the facility location problem. Before his doctoral studies, Professor Malladi worked as a Statistical Analyst for Fractal Analytics and as an Equity Research Analyst for Centrum Capital. He completed his B. Tech in Mechanical Engineering from IIT Bombay and has an MBA from IIM Ahmedabad