Let GG act on manifolds MM and NN (on the left, unless stated otherwise).

A f:M→Nf:M\to N is GG-equivariant if

f(g⋅x)=g⋅f(x)for all g∈G, x∈M.f(g\cdot x)=g\cdot f(x)\qquad\text{for all }g\in G,\ x\in M.

If instead MM and NN carry right actions, equivariance means f(x⋅g)=f(x)⋅gf(x\cdot g)=f(x)\cdot g for all gg.

Equivariant maps are precisely morphisms in the category of manifolds with a of GG. They preserve orbits and stabilizers (up to inclusion): f(G⋅x)⊆G⋅f(x)f(G\cdot x)\subseteq G\cdot f(x) and Gx⊆Gf(x)G_x\subseteq G_{f(x)}.

Examples
  1. Bundle projection. For a π:P→B\pi:P\to B with right action on PP and the trivial action on BB, the projection is equivariant (it is invariant): π(p⋅g)=π(p)\pi(p\cdot g)=\pi(p).
  2. Inversion under conjugation. For the conjugation action of GG on itself, the inversion map i(g)=g−1i(g)=g^{-1} is equivariant because i(hgh−1)=hg−1h−1i(hgh^{-1})=hg^{-1}h^{-1}.
  3. Intertwiners of representations. If GG acts linearly on vector spaces V,WV,W (representations), then a linear map T:V→WT:V\to W is equivariant exactly when T(ρV(g)v)=ρW(g)T(v)T(\rho_V(g)v)=\rho_W(g)T(v) for all g,vg,v.