Title: Visualization-Aided Multi-criterion Optimization and Decision-Making
Date and time: 11 March 2025 (Tuesday), 11:30 a.m. – 12:30 p.m.
Venue: IEOR Seminar Room
Abstract: Evolutionary algorithms (EAs) are effectively used to solve multi- and many-objective optimization problems to arrive at a set of non-dominated solutions. However, from an implementation point of view, decision makers (DMs) must pick one or a few solutions that satisfy their requirements or criteria. Multi-criteria decision-making (MCDM) techniques permit DMs to specify their criteria to obtain one or a few preferred solutions, iteratively. As an iterative method, MCDM techniques are usually accompanied by visualization techniques to assist DMs in effectively providing their criteria to arrive at and select the preferred solution(s) interactively. However, there are a few challenges in existing MCDM techniques: (1) Existing visualization techniques used in MCDM are effective in only a few dimensions, as they lose information and/or interpretability in higher (>3) dimensions. (2) Existing MCDM methods usually allow DMs to provide criteria only in the objective/design space. Often, a DM prefers to specify criteria based on trade-offs, closeness to constraint boundaries, and other factors. (3) The impact of small perturbations in decision variables on preferred solutions is rarely considered in the EA-based MCDM literature. These small perturbations often result from manufacturing tolerances, process parameters, or uncertain environmental conditions, and must be accounted for in the MCDM formulation. (4) An easy-to-use graphical user interface (GUI) is needed in interactive MCDM tasks to allow a DM to input their preferences, analyze and compare the current solutions using visualization techniques, and select a final solution. In this talk, I will present my current research on developing a visualization-assisted MCDM framework that addresses the aforementioned challenges in interactive MCDM literature.
Bio: Deepanshu is a Ph.D. graduate from the Department of Engineering Design at the Indian Institute of Technology (IIT) Madras, Tamil Nadu, India. He earned his B.Tech. in Mechanical Engineering from the National Institute of Technology (NIT) Kurukshetra, Haryana, India. Before joining IIT Madras through the Integrated M.S. + Ph.D. program, he gained industry experience at Vedanta Aluminium Limited, Odisha, India. He was also a visiting scholar at Michigan State University (MSU), USA. His Ph.D. research focuses on developing integrated frameworks for visualization-assisted multi-criteria decision-making. His research interests include data-driven modeling, optimization, and decision-making.