Adopting robust decision-making to forest management under climate change

Table 1 Robustness criteria used in the reviewed robust decision-making frameworks and related concepts

	Regret-based	Satisficing-based	Satisficing-optimizing
	Relative performance criteria	Absolute performance criteria
Goal	Minimizing the maximum deviation of an alternative’s performance from estimated/best performance across scenarios	Maximize the fraction of scenarios in which an alternative meets performance criteria	Minimize the worst-case performance
Metrics	Regret to best estimate scenario: candidate alternative’s maximum performance deviation between best-estimate scenario and each other scenario	Global satisficing: fraction of scenarios of candidate alternative in which performance requirements are met	All constraints to the objective function have to be fulfilled over the uncertainty set (scenarios)
	Regret to best performing alternative: maximum performance deviation of best-performing alternative and candidate alternative for each scenario	Local satisficing: number of uncertainty intervals from a best estimate outwards until candidate alternative fails performance requirements
Related concepts	Savage’s minimax regret	Radius of stability	Wald’s maximin

Both regret-based and satisficing-based robustness criteria have two metrics each to quantify robustness. For explanations of Savage’s minimax regret, Wald’s maximin and the radius of stability please refer to the text, Sect. 4

ISSN: 1297-966X