Piscopo, Amy N听1听;听Kasprzyk, Joseph R听2听;听Neupauer, Roseanna M听3
1听University of Colorado Boulder
2听University of Colorado Boulder
3听University of Colorado Boulder
The need for robust methods to optimize design problems is ubiquitous across many fields of engineering. Multi-objective evolutionary algorithms (MOEA) are used to optimize complex design problems using embedded simulation models that evaluate the performance of a population of candidate solutions in a single algorithm run. Solutions are evaluated based on objectives established by decision-makers, who also designate the decision variables and constraints of the design problem. Objectives, decision variables, and constraints constitute the problem formulation, a necessary and important component for optimization because it directly influences the results generated by the MOEA. Typically, design problems are optimized based on a single problem formulation established a priori; however, in this work, we demonstrate an approach to optimize iteratively using problem formulations updated from analyses of results from prior rounds of optimization. To demonstrate the approach, we consider a novel groundwater remediation technique, Engineered Injection and Extraction (EIE), which has never been optimized in the evolutionary algorithm literature. The problem is characterized by multiple conflicting objectives and uncertainty in our knowledge of subsurface properties and contaminant concentrations. Solutions generated by the MOEA using iterative problem reformulation have better performance than the baseline EIE sequence used in prior work.