Interactions between (dependent) functions and multilinear maps #
Main definitions #
MultilinearMap.pi_ext, a multilinear version ofLinearMap.pi_extMultilinearMap.piFamily, which satisfiespiFamily f x p = f p (fun i => x i (p i)).This is useful because all the intermediate results are bundled:
MultilinearMap.piFamily fis aMultilinearMapoperating on functionsx.MultilinearMap.piFamilyₗis aLinearMap, linear in the family of multilinear mapsf.
Two multilinear maps from finite families are equal if they agree on the generators.
This is a multilinear version of LinearMap.pi_ext.
Given a family of indices κ and a multilinear map f p for each way p to select one index from
each family, piFamily f maps a family of functions (one for each domain κ i) into a function
from each selection of indices (with domain Π i, κ i).
Equations
- MultilinearMap.piFamily f = { toFun := fun (x : (i : ι) → (j : κ i) → M i j) (p : (i : ι) → κ i) => (f p) fun (i : ι) => x i (p i), map_update_add' := ⋯, map_update_smul' := ⋯ }
Instances For
When applied to a family of finitely-supported functions each supported on a single element,
piFamily is itself supported on a single element, with value equal to the map f applied
at that point.
When only one member of the family of multilinear maps is nonzero, the result consists only of the component from that member.
MultilinearMap.piFamily as a linear map.
Equations
- MultilinearMap.piFamilyₗ = { toFun := MultilinearMap.piFamily, map_add' := ⋯, map_smul' := ⋯ }