You are here:
Integrative systems modeling and multi-objective optimization
Citation:
Barnhart, B. Integrative systems modeling and multi-objective optimization. SSWR Task 5.01A Meeting, Corvallis, OR, April 04, 2016.
Impact/Purpose:
Ecohydrological and economic models are commonly integrated to investigate the trade-offs and combined effects of human actions on systems-level outcomes, including environmental, economic, and human health objectives. This presentation demonstrates a number of applications of multi-objective optimization algorithms with integrated systems models to better understand the impacts of green infrastructure and agricultural best management practices on water quality and habitat integrity. A methodology to incorporate climate uncertainty into multi-objective optimization algorithms is also demonstrated. Finally, a new method for creating water quality indexes is demonstrated. This presentation provides a number of useful algorithms, tools, and methodologies that researchers can use within their own watershed-scale modeling studies. This work contributes to milestones within Product 3 of Task 5.01A in SSWR.
Description:
This presentation presents a number of algorithms, tools, and methods for utilizing multi-objective optimization within integrated systems modeling frameworks. We first present innovative methods using a genetic algorithm to optimally calibrate the VELMA and SWAT ecohydrological models on EPA’s SOL high-performance computing cluster. Also, a relatively new method for spatial allocation of agri-environmental policy in the Calapooia River basin is demonstrated by linking economic and biophysical models within a multi-objective optimization framework. Climate uncertainty within this optimization framework is discussed, and a method to incorporate the effects of uncertainty on optimal spatial allocation of policy is presented. Finally, a new index method for aggregating multiple nutrient constituents into a single time-varying water quality index is presented. The algorithms, tools, and methods presented in this work will be directly applicable to watershed-scale modelers who wish to utilize optimization methods within integrated systems modeling frameworks.