We consider the problem of control allocation for weakly redundant nonlinear plant models, that is, input redundant systems possessing a maximum rank control distribution. Attention is focused on nonlinear models that admit a special global normal form, which subsumes the case of systems with globally-defined vector relative degree. It is shown how the availability of a set of suitably defined redundant input leads (under some technical conditions) to the definition of different static and dynamic strategies for steady-state optimization within a given set-point regulation problem