Integrating Interactive Decision Making Into Evolutionary Multiobjective Agricultural Optimization
ORCID
Kropp: https://orcid.org/0000-0002-3779-2801; Jha: https://orcid.org/0000-0001-5973-711X; Pulido: https://orcid.org/0000-0001-5311-8027
MSU Affiliation
College of Agriculture and Life Sciences; Department of Plant and Soil Sciences
Creation Date
2026-04-29
Abstract
Humanity must increase agricultural output to feed its increasingly large and often affluent population. Currently, industrialized agriculture can meet these demands, but it is at the expense of water quantity and quality. Existing research has successfully applied evolutionary multiobjective optimization (EMO) to optimize field-scale irrigation schedules, significantly reducing water use while maintaining yield. However, a lack of trust between farmers and optimization software stands in the way of adoption of such tools, and the optimization runs infeasibly slow. In this research, we apply progressively-interactive EMO (PI-EMO) to incorporate human decision-making into the optimization process to make the algorithm more trustworthy and faster. To do so, we implemented an improved publicly available version of the PI-NSGA-II algorithm for use in irrigation management. We then compared the performance of NSGA-II (a non-interactive EMO) and PI-NSGA-II against 900 different irrigation optimization scenarios. Across these scenarios, the overall performance of PI-NSGA-II was comparable to that of NSGA-II while running significantly faster with half the function evaluation. In doing so, we were able to: (1) successfully incorporate interactivity into the agricultural optimization problem, and (2) reduce the number of function evaluations in half. This had a median runtime improvement of 23% between the two algorithms. In other words, PI-EMO-VF makes EMO agricultural optimization more trustworthy and faster than current state-of-the-art EMO irrigation platforms. This study demonstrates the feasibility of future optimization decision support tools, which will empower farmers to make decisions that are not only productive, but maintain the agro-ecosystem supporting humanity.
Publication Date
12-24-2025
Publication Title
Computers and Electronics in Agriculture
Publisher
Elsevier
Recommended Citation
Kropp, I., Nejadhashemi, A. P., Jha, P. K., Spinner, E., & Pulido, G. T. (2026). Integrating interactive decision making into evolutionary multiobjective agricultural optimization. Computers and Electronics in Agriculture, 242, 111359. https://doi.org/10.1016/j.compag.2025.111359
