Stochastic Modeling with Phase-type Distributions Introduction and Some Applications
School of Management, Sabancı University
Phase-type is a family of continuous distributions based on the method of stages. They model the random variables as being composed of a number of exponentially distributed phases, exploiting the resulting Markovian structure that facilitates the analysis of many stochastic systems. These distributions possess some very useful properties for stochastic modeling: their distribution function, expected value, and higher-order moments can be found in closed form. They are closed under convolution and can approximate any positive, continuous distribution or data trace with arbitrary precision. In this talk, I first introduce the structure and properties of Phase-type distributions. Then I showcase their application in a variety of stochastic models applied to revenue management, vehicle routing, infrastructure reliability and healthcare.
Raha Akhavan-Tabatabaei is associate professor of Operations Management and co-director of Masters in Business Analytics at Sabanci School of Management. Prior to this position, she was associate professor of Industrial Engineering and the founding director of Masters in Analytics at Universidad de los Andes in Bogota, Colombia. Before that, she worked as senior industrial engineer at Intel Corporation in Chandler, Arizona. She has received her Ph.D. and M.Sc. in Industrial Engineering and Operations Research from North Carolina State University, and her B.Sc. from Sharif University of Technology. Her research is focused on stochastic modeling and data-driven decision making with applications in healthcare, logistics, revenue management and reliability.
Friday, November 10, 2017 at 4.00 pm in IE03