Bonham, N., Kasprzyk, J. and Zagona, E. (2022). "post-MORDM: Mapping Policies to Synthesize Optimization and Robustness Results for Decision-maker Compromise," Environmental Modelling & Software, 157, November 2022, 105491, (https://doi.org/10.1016/j.envsoft.2022.105491).
Abstract
This paper introduces post-MORDM, a decision-support framework that augments Many Objective Robust Decision Making (MORDM). MORDM often creates an intractable number of environmental management policies, characterized by decision variable, objective, and robustness values. This large number of policies inhibits decision support, causing disagreements among decision-makers. Post-MORDM addresses these challenges via the Self-Organizing Map (SOM), synthesizing MORDM data as layers organized in a map-like coordinate system. It uses the SOM to cluster policies, discover salient characteristics, and assess cause-effect relationships between decision-maker choices (i.e. decision variable values) and performance (objective and robustness values). Overall, the goal of post-MORDM is to create a structured platform that encourages negotiation and compromise. We demonstrate post-MORDM with a case study of two illustrative decision-makers for reservoir operation policy in the Colorado River Basin, USA. Post-MORDM helps communicate tradeoffs between storage and delivery objectives, relate tradeoffs to shortage policies, and identify mutually feasible policies.
Graphical Abstract
Download higher-resolution version of the above image
Keywords: Self-organizing maps (SOM); Decision-support; Robustness; Tradeoff analysis; Colorado River Basin; Multi-objective evolutionary algorithm (MOEA)