04a Ito Rules

$$ \newcommand{aster}{*} \newcommand{exist}{\exists} \newcommand{B}{\mathbb B} \newcommand{C}{\mathbb C} \newcommand{I}{\mathbb I} \newcommand{N}{\mathbb N} \newcommand{Q}{\mathbb Q} \newcommand{R}{\mathbb R} \newcommand{Z}{\mathbb Z} \newcommand{eR}{\overline {\mathbb R}} \newcommand{cD}{ {\mathbb D}} \newcommand{dD}{ {\part \mathbb D}} \newcommand{dH}{ {\part \mathbb H}} \newcommand{eC}{\overline {\mathbb C}} \newcommand{A}{\mathcal A} \newcommand{D}{\mathcal D} \newcommand{E}{\mathcal E} \newcommand{F}{\mathcal F} \newcommand{G}{\mathcal G} \newcommand{H}{\mathcal H} \newcommand{J}{\mathcal J} \newcommand{L}{\mathcal L} \newcommand{U}{\mathcal U} \newcommand{M}{\mathcal M} \newcommand{O}{\mathcal O} \newcommand{P}{\mathcal P} \newcommand{S}{\mathcal S} \newcommand{T}{\mathcal T} \newcommand{V}{\mathcal V} \newcommand{W}{\mathcal W} \newcommand{X}{\mathcal X} \newcommand{Y}{\mathcal Y} \newcommand{bE}{\symbf E} \newcommand{bF}{\symbf F} \newcommand{bD}{\symbf D} \newcommand{bI}{\symbf I} \newcommand{bX}{\symbf X} \newcommand{bY}{\symbf Y} \newcommand{nz}{\mathcal Z} \newcommand{bT}{\mathbb T} \newcommand{bB}{\mathbb B} \newcommand{bS}{\mathbb S} \newcommand{bA}{\mathbb A} \newcommand{bL}{\mathbb L} \newcommand{bP}{\symbf P} \newcommand{bM}{\symbf M} \newcommand{bH}{\mathbb H} \newcommand{dd}{\mathrm d} \newcommand{Mu}{\mathup M} \newcommand{Tau}{\mathup T} \newcommand{ae}{\operatorname{a.e.}} \newcommand{aut}{\operatorname{aut}} \newcommand{adj}{\operatorname{adj}} \newcommand{char}{\operatorname{char}} \newcommand{cov}{\operatorname{Cov}} \newcommand{cl}{\operatorname{cl}} \newcommand{cont}{\operatorname{cont}} \newcommand{e}{\mathbb E} \newcommand{pp}{\operatorname{primitive}} \newcommand{dist}{\operatorname{dist}} \newcommand{diam}{\operatorname{diam}} \newcommand{fp}{\operatorname{Fp}} \newcommand{from}{\leftarrow} \newcommand{Gal}{\operatorname{Gal}} \newcommand{GCD}{\operatorname{GCD}} \newcommand{LCM}{\operatorname{LCM}} \newcommand{fg}{\mathrm{fg}} \newcommand{gf}{\mathrm{gf}} \newcommand{im}{\operatorname{Im}} \newcommand{image}{\operatorname{image}} \newcommand{inj}{\hookrightarrow} \newcommand{irr}{\operatorname{irr}} \newcommand{lcm}{\operatorname{lcm}} \newcommand{ltrieq}{\mathrel{\unlhd}} \newcommand{ltri}{\mathrel{\lhd}} \newcommand{loc}{ {\operatorname{loc}}} \newcommand{null}{\operatorname{null}} \newcommand{part}{\partial} \newcommand{pf}{\operatorname{Pf}} \newcommand{pv}{\operatorname{Pv}} \newcommand{rank}{\operatorname{rank}} \newcommand{range}{\operatorname{range}} \newcommand{re}{\operatorname{Re}} \newcommand{span}{\operatorname{span}} \newcommand{su}{\operatorname{supp}} \newcommand{sgn}{\operatorname{sgn}} \newcommand{syn}{\operatorname{syn}} \newcommand{var}{\operatorname{Var}} \newcommand{res}{\operatorname{Res}} \newcommand{data}{\operatorname{data}} \newcommand{erfc}{\operatorname{erfc}} \newcommand{erfcx}{\operatorname{erfcx}} \newcommand{tr}{\operatorname{tr}} \newcommand{col}{\operatorname{Col}} \newcommand{row}{\operatorname{Row}} \newcommand{sol}{\operatorname{Sol}} \newcommand{lub}{\operatorname{lub}} \newcommand{glb}{\operatorname{glb}} \newcommand{ltrieq}{\mathrel{\unlhd}} \newcommand{ltri}{\mathrel{\lhd}} \newcommand{lr}{\leftrightarrow} \newcommand{phat}{^\widehat{\,\,\,}} \newcommand{what}{\widehat} \newcommand{wbar}{\overline} \newcommand{wtilde}{\widetilde} \newcommand{iid}{\operatorname{i.i.d.}} \newcommand{Exp}{\operatorname{Exp}} \newcommand{abs}[1]{\left| {#1}\right|} \newcommand{d}[2]{D_{\text{KL}}\left (#1\middle\| #2\right)} \newcommand{n}[1]{\|#1\|} \newcommand{norm}[1]{\left\|{#1}\right\|} \newcommand{pd}[2]{\left \langle {#1},{#2} \right \rangle} \newcommand{argmax}[1]{\underset{#1}{\operatorname{argmax}}} \newcommand{argmin}[1]{\underset{#1}{\operatorname{argmin}}} \newcommand{p}[1]{\left({#1}\right)} \newcommand{c}[1]{\left \{ {#1}\right\}} \newcommand{s}[1]{\left [{#1}\right]} \newcommand{a}[1]{\left \langle{#1}\right\rangle} \newcommand{cc}[2]{\left(\begin{array}{c} #1 \\ #2 \end{array}\right)} \newcommand{f}{\mathfrak F} \newcommand{fi}{\mathfrak F^{-1}} \newcommand{Fi}{\mathcal F^{-1}} \newcommand{l}{\mathfrak L} \newcommand{li}{\mathfrak L^{-1}} \newcommand{Li}{\mathcal L^{-1}} \newcommand{const}{\text{const.}} $$

Itô Rules in 1D #

Itô processes #

Suppose $(\Omega, \F, (\F _ t), [0, T], P)$ is a complete filtered probability space. Let $(B _ t)$ be a standard Brownian motion.

Suppose for non-anticipating $(F _ t) \in \bA _ 1(0, T)$ and $(G _ t) \in \bA _ 2(0, T)$, $(X _ t)$ is a real-valued stochastic process satisfying $$ \forall 0 \le s \le r \le T: P\p{X _ r = X _ s + \int _ s^r F _ t \dd t + \int _ s^r G _ t \dd B _ t} = 1, $$ we say that $(X _ t)$ is an 1-dimensional Itô process and write $$ \dd X _ t = F _ t \dd t + G _ t \dd B _ t, \quad t \in [0, T] $$

  • Notice that $(X _ t)$ has continuous paths almost surely.
  • $(G _ t) \in \bA^2(0, T)$ guarantees that $\p{\int _ 0^t G _ s \dd B _ s} \in \bL _ {c, [2], m}^2(0, T)$.
  • $(F _ t) \in \bA^1(0, T)$ guarantees that $\p{\int _ 0^t F _ s \dd s} \in \bL _ {c, [1]}^1(0, T)$.
  • When $P(X _ 0 = 0) = 1$ the process is said to have a zero start.
  • $(G _ t)$ is called the local volatility, $(F _ t)$ is called the local drift.
Basic transforms of the Brownian motion #

Suppose $(\Omega, \F, (\F _ t), [0, T], P)$ is a complete filtered probability space. Let $(B _ t)$ be a standard Brownian motion.

We can demonstrate that $$ \dd B _ T^2 = \dd t + 2B _ t \dd B _ t, \quad t \in [0, T] $$

  • Given partition $(t _ {k}) _ {k = 0}^{m _ n} \in P[s, t]$ we have $$ \sum _ {k = 0}^{m _ n - 1} B _ {t _ k}(B _ {t _ {k + 1}} - B _ {t _ k}) = \frac{B _ t^2 - B _ s^2}{2} - \frac{1}{2} \sum _ {k = 0}^{m _ n - 1} (B _ {t _ {k + 1}} - B _ {t _ k})^2 $$
  • Now take any sequence of shrinking partition $\p{(t _ {n, k}) _ {k = 0}^{m _ n}} _ {n = 1}^\infty \subseteq P[s, t]$. We have that with probability one: $$ \int _ s^t B _ \tau \dd B _ \tau = \frac{B _ t^2 - B _ s^2}{2} + \frac{t - s}{2} $$

Also we have $$ \dd (t B _ t) = B _ t \dd t + t \dd B _ t,\quad t \in [0, T] $$

  • Given partition $(t _ {k}) _ {k = 0}^{m _ n} \in P[s, t]$ we have $$ \sum _ {k = 0}^{m _ n - 1} t _ {k} (B _ {t _ {k + 1}} - B _ {t _ k}) + \sum _ {k = 0}^{m _ n - 1} B _ {t _ {k + 1}}(t _ {k + 1} - t _ k) = t B _ t - sB _ s $$
  • Now take any sequence of shrinking partition $\p{(t _ {n, k}) _ {k = 0}^{m _ n}} _ {n = 1}^\infty \subseteq P[s, t]$. We have that with probability one: $$ t B _ t - sB _ s = \int _ s^t B _ \tau \dd \tau + \int _ s^t \tau \dd B _ \tau $$
Itô chain rule TODO #

Suppose $(\Omega, \F, (\F _ t), [0, T], P)$ is a complete filtered probability space. Let $(B _ t)$ be a standard Brownian motion. Suppose $$ \dd X _ t = F _ t \dd t + G _ t \dd B _ t,\quad t \in [0, T] $$ Suppose $u(x, t) \in C^{2, 1}(\R \times [0, T] \to \R)$, that is $D _ 1 u, D _ {11} u, D _ 2 u$ all exist and are continuous.

  • For convenience, we denote $u:= u(X _ t, t)$, $u _ x := D _ 1 u(X _ t, t)$, $u _ t := D _ 2 u(X _ t, t)$, $u _ {xx} := D _ {11}u(X _ t, t)$ in the following discussion.

Then $Y _ t = u(X _ t, t)$ has stochastic differential $$ \dd Y _ t = \dd u(X _ t, t) = u _ t \dd t + u _ x \dd X _ t + \frac{1}{2} u _ {xx} G _ t^2 \dd t = \p{u _ t + u _ x F _ t + \frac{1}{2} u _ {xx} G _ t^2} \dd t + u _ x G _ t \dd B _ t, \quad t \in [0, T] $$

  • We always have $\p{u _ t + u _ x F _ t + \frac{1}{2} u _ {xx} G _ t^2} \in \bA _ {1}(0, T)$ and $(u _ x G _ t) \in \bA _ {2}(0, T)$, they are both non-anticipating.
Generalized Itô chain rule TODO #

Suppose $(\Omega, \F, (\F _ t), [0, T], P)$ is a complete filtered probability space. Let $(B _ t)$ be a standard Brownian motion. Suppose $$ \forall i \in \c{1, \ldots, n}:\dd X _ t^{(i)} = F _ t^{(i)} \dd t + G _ t^{(i)} \dd B _ t, \quad t \in [0, T] $$ Suppose $u \in C^{1, 2}(\R^n \times [0, T] \to \R)$. Then let $u _ t := D _ t u(\symbf X _ t, t)$, $u _ i:= D _ {x _ i} u(\symbf X _ t, t)$ and so on. $$ \dd u(\symbf X _ t, t) = u _ t \dd t + \sum _ {i = 1}^{n} u _ {i} \dd X^{(i)} _ t + \frac{1}{2} \sum _ {i, j = 1}^n u _ {ij} G _ t^{(i)} G _ t^{(j)} \dd t,\quad t \in [0, T] $$

Itô's product rule #

Suppose $(\Omega, \F, (\F _ t), [0, T], P)$ is a complete filtered probability space. Let $(B _ t)$ be a standard Brownian motion.

And suppose $$ \dd X _ t = G _ t \dd t + U _ t \dd B _ t, \quad \dd Y _ t = H _ t \dd t + V _ t \dd B _ t, \quad t \in [0, T] $$ Then immediately $$ \begin{aligned} \dd (X _ t + Y _ t) & = (G _ t + H _ t) \dd t + (V _ t + U _ t) \dd B _ t,& t \in [0, T]\\ \dd (X _ t Y _ t) & = \frac{1}{2} \p{\dd (X _ t + Y _ t)^2 - \dd X _ t^2 - \dd Y _ t^2}, & t \in [0, T] \end{aligned} $$ Consider $u(x, t) = x^2$, we have $$ \begin{aligned} &\dd X _ t^2 = 2X _ t \dd X _ t + U _ t^2\dd t = 2X _ t (G _ t \dd t + U _ t \dd B _ t) + U _ t^2\dd t = (2X _ t G _ t + U _ t^2) \dd t + 2X _ t U _ t \dd B _ t\\ &\dd Y _ t^2 = 2Y _ t \dd Y _ t + V _ t^2 \dd t = 2Y _ t (H _ t \dd t + V _ t \dd B _ t) + V _ t^2\dd t = (2Y _ t H _ t + V _ t^2) \dd t + 2Y _ t V _ t \dd B _ t\\ &\dd(X _ t + Y _ t)^2 = 2(X _ t+ Y _ t) \dd (X _ t + Y _ t) + (U _ t + V _ t)^2 \dd t\\ &\dd (X _ t + Y _ t)^2 = \p{2(X _ t + Y _ t)(H _ t + G _ t) + (U _ t + V _ t)^2} \dd t + 2(X _ t + Y _ t)(U _ t+ V _ t) \dd B _ t \end{aligned} $$ Therefore by subtraction, $$ \begin{aligned} \dd (X _ t Y _ t) &= X _ t \dd Y _ t + Y _ t \dd X _ t + U _ t V _ t \dd t\\ &= X _ t(H _ t \dd t + V _ t \dd B _ t) + Y _ t (G _ t \dd t + U _ t \dd B _ t) + U _ t V _ t \dd t\\ & = (X _ t H _ t + Y _ t G _ t + U _ t V _ t) \dd t + (X _ t V _ t + Y _ t U _ t) \dd B _ t \end{aligned} $$

Notice that the same result also follows from the generalized Itô's chain rule.