Visualization-based multi-objective improvement of environmental decision-making using linearization of response surfaces

This study presents a new interactive procedure for supporting Decision Makers (DMs) in environmental planning problems involving large, process-based, dynamic models and many (more than two) conflicting objectives. Because of such features of the model, computationally-onerous simulations are the only feasible way of analysis, while the multi-objective nature of the problem entails the combined use of optimization techniques and appropriate tools for the visualization of the associated Pareto frontier. The procedure proposed is based on the iterative improvement of the current best compromise alternative based on interactions with the DM. At each iteration, the DM is informed about the Pareto frontier of a local multi-objective optimization problem, which is generated by linearizing the response surfaces that describe the objectives and constraints of the original planning problem. Interactive visualization of the multi-dimensional Pareto frontier is used to support the DM in choosing the new best compromise alternative. The procedure terminates when the DM is fully satisfied with the current best compromise alternative. The approach is demonstrated in Googong Reservoir (Australia), which is periodically affected by high concentrations of Manganese and Cyanobacteria. Results indicate that substantial improvements could be observed by simply changing the location of the two mixers installed in 2007 and adding another pair of mixers.