Stochastic calculus with infinitesimals (Q455364)

This slim volume of lecture notes provides an elegant and intuitive introduction to stochastic analysis with applications from an original but highly natural viewpoint. Its perspective of a slightly extended (nonstandard) model of the naturals and reals, the so-called minimal internal set theory, which includes a rigorous notion of infinitesimals, allows to prove many classical theorems of stochastic analysis by means of huge but finite (nonstandard) probability spaces and, as a consequence, without any measure theory or functional analysis. This approach goes back to the work of \textit{E. Nelson} [Radically elementary probability theory. Princeton University Press (1987; Zbl 0651.60001)], who presented an introduction to probability theory based on the concept of internal set theory, a conservative extension of standard analysis and an alternative concept to Robinson's nonstandard analysis. In fact, the author introduces a substantially smaller set of axioms, the so-called minimal internal set theory, which allows for a theory of stochastic processes in continuous time. The material is organized in ten chapters and two appendices about the model theoretic context of this setting. The author starts with a short and elementary introduction to minimal internal set theory, the very basis of this text. Chapter~2 is devoted to a swift introduction to radically elementary probability theory including the construction of the Wiener and the Poisson walk, which are the equivalent to the standard Wiener and Poisson process. In Chapter 3, the author lays out a radically elementary (Itō) integration theory without any measure theory. It starts with the (\(L^2\)-) martingale representation theorem and the Itō isometry followed by the essential uniqueness of the Itō decomposition of Itō processes, the Itō-Doeblin change-of-variable formula, and closes with the positivity and integrability of the geometric Itō processes (with monotone drift). The chapter is completed with the statement of Nelson's version of Lévy's characterization of Wiener processes. A highlight of the text is arguably the radically elementary version of Girsanov's change-of-measure theorem and the diffusion invariance principle in Chapter 4. The next three chapters are devoted to applications. Chapter 5 provides a neat introduction to financial economics in this setting. Here, the nonstandard approach shows its full strength by allowing a clear and intuitive proof of the fundamental theorem of asset pricing in less than three pages. This is followed by a discussion of the Black-Scholes model and its volatility invariance in Chapter 6. The next excursion leads to mathematical physics. Chapter 7.1 deals with Itō's formula for time-homogeneous Itō diffusions, which is the basis for a proof of the Markov property of Itō diffusions, the Feynman-Kac formula, a radically elementary definition of Feyman integrals and a rigorous definition of the Feynman path integral-based on minimal internal set theory in Chapter 8. Chapter 9 provides a radically elementary introduction to the theory of Lévy processes. It starts with random and Lévy walks and the existence of exponential moments for Lévy walks with bounded increments. It gives an interesting criterion for random walks to be Lévy walks and culminates with the Itō-Doeblin formula for Lévy walks with limited variation jump-part. Eventually, the author gives a very short introduction to Lévy finance. The radically elementary setting of minimal internal set theory puts the author in the position to prove many important theorems of stochastic calculus on a few elegant pages and with nothing more but elementary though nonstandard reasoning. The approach merits to be still pushed further, for instance towards a radically elementary Malliavin's calculus. It is particularly suitable for curricula of quantitative finance, engineering and physics, as the author suggests, but due to the economy of its reasoning by far not limited to that.