Ersatzmodell

Ersatzmodell is a term used mainly in German-speaking contexts to refer to a substitute or surrogate model used in analysis, design, or testing when the real system or ideal model is unavailable, impractical, or too costly to use. The word ersatz means "substitute" or "replacement." An ersatzmodell can be physical—a smaller, simplified version of an object or system, such as a scale model or mock-up—or mathematical/computational—a surrogate model that approximates the behavior of a more complex model or of the real system.

In practice, ersatzmodelle are used to explore hypotheses, perform preliminary testing, or guide decision-making. They enable rapid experiments, sensitivity analysis, optimization, and training in simulation environments without incurring the expense or risk of using the actual system. Common forms include surrogate models in statistical learning (e.g., response surfaces, kriging, neural networks), reduced-order models in engineering, and physical mock-ups in architecture or product design.

A key distinction is that an ersatzmodell trades accuracy for tractability; its validity depends on the domain, the quality of the approximation, and the context of use. When the real model or system becomes available, results should be validated against it, and the ersatzmodell may be replaced or updated.

Ersatzmodelle are related to terms such as surrogate model, proxy model, or mock-up.