No label defined (Q4802403)

The authors give an explicit construction for the vector stochastic integral \(H\cdot X\) of a \(d\)-dimensional process \(H\) with respect to a \(d\)-dimensional semimartingale \(X\) that takes into account the interference of different components of the process \(X\) and allows to deal with integrands \(H\) which are not locally bounded. In particular, given a \(d\)-dimensional local martingale \(M\) defined on the probability space \((\Omega, { F}, P)\) and a simple integrand \(H\) of the form \[ H_t=h_0I(t=0)+\sum_{k=1}^{d}h_kI(\tau_k\leq t\leq \tau_{k+1}),\quad t\geq 0, \] they define the integral \(H\circ M\) as \[ (H\circ M)_t= \sum_{i=1}^{d}\sum_{k=1}^{m}h^i_k(M^i_{t\wedge \tau_{k+1}}- M^i_{t\wedge \tau_{k}}). \] Here \(0\leq \tau_1\leq \dots\leq \tau_{n+1}\) are stopping times and \(h_k\in R^d\) are bounded and \({ F}_{\tau_k}\) measurable. Furthermore the authors prove that the space of vector stochastic integrals is closed in the appropriate topology. In the final part of the article the authors show that the notion of a vector stochastic integral is important in mathematical finance. Among others it is proved to be crucial in the proof of the first and second fundamental theorems of asset pricing. For the entire collection see [Zbl 1007.00020].