Title: Lending insight to environmental problems characterized by conflicting objectives

Abstract: In environmental planning problems, decision makers' goals often fundamentally conflict.  Faced 
with limited budgets, planners realize that increasing their systems' performance could cause increased 
costs, negative impacts to the environment, or conflicts among stakeholders.  Multi-objective decision 
support techniques are increasingly used to generate planning alternatives for such planning problems, with 
the idea that exposing multiple candidate alternative solutions can positively contribute to the planning 
process.  For example, we have investigated a multi-objective decision support process for water supply 
planning in the Tarrant Regional Water District of Texas, finding better solutions to balance multiple 
supply reservoirs; and for the design of groundwater contamination strategies, gaining increased contaminant 
degradation with reduced pumping rates.  Our research group uses multi-objective evolutionary algorithm 
search, simulation models, interactive visualizations, and high performance computing to solve 
multi-objective decision support in a number of different domains that span civil and environmental 
engineering.

This presentation will first introduce the methodology for multi-objective decision support.  The remainder 
of the presentation will explore several key areas where multi-objective approaches can actually add insight 
to our understanding of environmental problems, in addition to providing new planning alternatives.  First, 
a series of decision support experiments will show that changes to the optimization problem formulation 
(decision variables, objectives, and constraints) fundamentally alters the solution algorithms' ability to 
find answers to decision problems.  Second, we will show how our approach can contribute to the creation of 
new environmental regulations, using an example from unconventional oil and gas development in Colorado.  
Finally, we will discuss how our suite of tools can help aid in performance assessment of quantitative 
hydrological models such as conceptual and physically-based snow models.