message_ix_models.model.transport.operator.logit

message_ix_models.model.transport.operator.logit(x: AnyQuantity, k: AnyQuantity, lamda: AnyQuantity, y: list[int], dim: str) → AnyQuantity

Compute probabilities for a logit random utility model.

The choice probabilities have the form:

$$Pr(i)=\frac{k_j{x}_j^{{\lambda }_j}}{\sum _{\mathrm{\forall }i\in D}{k}_i{x}_i^{{\lambda }_i}}\mathrm{\forall }j\in D$$

…where $D$ is the dimension named by the dim argument. All other dimensions are broadcast automatically.