Notion of quadratic variation in Banach spaces

This article introduces the notion of global quadratic variation which is based on a notion weaker than the usual tensor topology appearing in the literature of stochastic processes in Banach separable spaces. It is based on convergence in the weak-star topology of approximating sequences. This calculus pursues in the infinite dimensional space the stochastic calculus via regularization introduced by Russo and Vallois for real valued processes and developed in Banach space by Di Girolami and Russo. This article in particular focuses on the relation with classical calculus in order to compare our global quadratic variation with the classical existing notions of quadratic variations in the literature.