Development of Metrics for Multiobjective Optimization of Environmental Remediation Problems

Michigan Technological University, Houghton MI

Investigators

Abstract

0083112 Mayer This research project focuses on the development of metrics for comparing various evolutionary multiobjective optimization methods that may be applied to the design of environmental systems, as part of a larger study for developing decision-making tools for environmental remediation. The main objective of this project is to advance the knowledge on evolutionary multiobjective optimization by (a) assessing the most common metrics currently in use, (b) developing new metrics, (c) conducting a comprehensive study of metrics in a real-world application and (d) conducting a sensitivity analysis of the existing and newly proposed metrics. The research developed as part of this project will make contributions to both computer science and operations research in the areas of evolutionary algorithms and multiobjective optimization, and will have important implications in decision making for environmental remediation. This project involves collaboration of the PI with Dr. Carlos Coello at the National Laboratory for Advanced Informatics (LANIA, Xalapa, Mexico) and Dr. Carlos Mariano at the Mexican Institute for Water Technology (IMTA, Cuernavaca, Mexico). The collaboration will take the form of communication via internet, biannual meetings of the investigators and the exchange of graduate students. This project will lay the groundwork for a long-term collaboration between Michigan Technological University, LANIA and IMTA. ***