► DEXi-like models are monotonic. ► Genetic algorithms can be used to find the best scenarios of DEXi-like models. ► The set of worst-best and best-worst scenarios is quite small when applied to the MASC model. ► The basic attributes with the highest sensitivity index are the most influential. DEXi-type models have been used recently to assess specific problems in agricultural systems and to assess cropping-system scenarios. Finding the set of the “worst-best” (lowest scores for basic attributes that lead to the highest score for the root attribute) and “best-worst” (highest scores for basic attributes that lead to the lowest score for the root attribute) options are of interest for improving current cropping systems. As DEXi-type models revealed a monotonicity property, we used a genetic algorithm to find these two sets. These sets are small and show that only a few attributes need to have low scores to reach the best-worst options or high scores to reach the worst-best options. These attributes are those with a high sensitivity index.