A decision-oriented framework for optimizing herbicide mixture ratios and dose bands under heterogeneous weed...

Herbicide mixtures are widely used in post-emergence weed management to broaden control spectra and mitigate resistance evolution. However, mixture selection is often based on empirical ratios or average efficacy metrics that do not adequately account for heterogeneous weed communities. In practice, management failure is typically driven by the most tolerant species rather than the community's mean performance. Despite this, quantitative decision criteria that explicitly account for worst-case efficacy remain limited.

A multi-stage decision-oriented framework was implemented in a sugarcane production system to optimize a ternary post-emergence herbicide mixture. The framework integrated ratio-dose response-surface modelling, interaction robustness screening, and conservative effective-dose mapping. Interaction analysis showed early positive deviations from additivity, which converged toward additivity at later assessments, indicating temporal dependence and limited persistence of synergistic effects. A conservative worst-case effective dose (ED90_worst), defined as the maximum ED90 across weed groups and assessment times, was used as the decision criterion for dose selection. At 25 days after application, ED90_worst ranged from 148.66 to 807.79 g a.i. ha^-1 across mixture ratios. The 30:8 ratio achieved ED90 for five of six weed groups. A practical dose band of 652.22-717.44 g a.i. ha^-1 satisfied worst-case efficacy requirements while maintaining crop safety across cultivars.

By prioritizing robustness and worst-case performance rather than point-optimal efficacy, this framework provides a reproducible decision-support approach for optimizing herbicide mixtures in heterogeneous weed communities and supports resistance-aware post-emergence weed management. © 2026 Society of Chemical Industry.