Zbigniew S. Kowalski
Faculty of Pure and Applied Mathematics, Wroclaw University of Science and Technology, Wroclaw, Poland.
DOI: 10.4236/tel.2024.146126   PDF    HTML   XML   127 Downloads   548 Views   Citations

Abstract

We show the Bernoulli property of the skew product transformation describing particle motion in a one-dimensional quantum harmonic oscillator. Finally, the application of the binomial model for asset prices is presented.

Keywords

Quantum Harmonic Oscillator, Step Skew Product Transformation, Bernoulli Property, Asset Prices

Share and Cite:

1. Introduction

A model of the motion of particles in a quantum harmonic oscillator on $ℝ$ was proposed (Kowalski, 2022). This is a skew-product transformation ${\widehat{S}}_p$ where $p\in \left(\frac{1}{2},\frac{1}{\sqrt[3]{2}}\right).$ The aim of the work is to show that ${\widehat{S}}_p$ has the Bernoulli property. Therefore, we will prove the ergodicity of ${\widehat{S}}_p$ . We consider its isomorphic version $S_p$ which is a random walk on $I=\left[0,1\right]$ . Firstly, we will describe the probability space on which $S_p$ operates. It is a product space $\left(\Omega \times I,ℬ\times \mathcal{A},{\mu }_p\times {\mu }_F\right)$ where $\Omega$ is the space ${\left\{0,1\right\}}^{\mathbb{N}}$ , $\mathbb{N}=\left\{0,1,2,\cdots \right\}$ , with the $\left(p,1-p\right)$ -Bernoulli measure ${\mu }_p$ on $\left(\Omega ,ℬ\right)$ . Here $ℬ$ is the Borel product $\sigma$ -algebra and $\mathcal{A}$ the Borel $\sigma$ -algebra of subsets of $ℝ$ . The measure ${\mu }_F$ has a probability distribution $F$ which is given by the mathematical description of the oscillator in (Kowalski, 2022) as follows. One-dimensional quantum harmonic oscillator $\Psi$ satisfies the Schrödinger equation

$i\hslash \frac{\partial \Psi }{\partial t}=H\Psi$

where $H=\frac{1}{2}\left(P^2+Q^2\right)$ is the Hamiltonian. Here $P=-i\frac{\text{d}}{\text{d}x}$ and $Q$ is multi-plication by $x$ . Let us consider the $n$ -quantum state solution

${\Psi }_n\left(x,t\right)=\frac{1}{\sqrt{n!2^n\sqrt{\pi }}}{H}_n\left(x\right)\exp \left(-\frac{i}{2\hslash }\left(2n+1\right)t\right)\exp \left(-\frac{x^2}{2}\right).$

for $n\ge 0$ . Here $H_n$ is $n$ th Hermite polynomial i.e.

$H_n\left(x\right)={\left(-1\right)}^n{\text{e}}^{x^2}\frac{{\text{d}}^n}{\text{d}{x}^n}{\text{e}}^{-x^2}.$

The quantum interpretation of

${\Phi }_n\left(x\right)={\displaystyle {\int }_{-\infty }^x{\left|{\Psi }_n\left(y,t\right)\right|}^2\text{d}y}=\frac{1}{n!2^n\sqrt{\pi }}{\displaystyle {\int }_{-\infty }^x{H}_n^2\left(y\right)\exp \left(-y^2\right)\text{d}y}$

is the probability distribution of occurrence of a particle in $ℝ$ . In Theorem 4.1 (Kowalski, 2017), it has been observed that for $n\ge 1$ there is a unique partition on intervals $\left\{I_k:k=1,\cdots ,2n\right\}$ of $ℝ$ such that ${\mu }_n\left(I_k\right)={\mu }_{n-1}\left(I_k\right)$ where the measure ${\mu }_n$ has distribution ${\Phi }_n$ . The endpoints of intervals come from the equivalence

${\Phi }_n\left(x\right)={\Phi }_{n-1}\left(x\right)\iff H_n\left(x\right)H_{n-1}\left(x\right)=0.$

It is convenient to consider the unit interval instead of $ℝ$ , therefore, we use the map ${\Phi }_{n-1}:ℝ\to \left(0,1\right)$ . ${\Phi }_{n-1}$ arise as the distribution function of the Lebesgue measure on the unit interval and ${\Phi }_n$ as $F^{\left(n\right)}={\Phi }_n\circ {\Phi }_{n-1}^{-1}$ . Here ${\Phi }_{n-1}^{-1}$ is the inverse function of ${\Phi }_{n-1}$ . We will denote $F^{\left(n\right)}$ by $F$ as $n$ is fixed. Next, we will describe the construction of the skew-product $S_p$ . Let $\sigma$ be the one-sided shift on $\Omega$ i.e. $\sigma \left(\omega \right)\left(i\right)=\omega \left(i+1\right)$ . Let us assume that conditions of Remark 2.6 (Kowalski, 2022) hold for some $p\in \left(\frac{1}{2},\frac{1}{\sqrt[3]{2}}\right]$ and for some $1\le k\le 2n-1$ . Then we get the step skew product transformation in the space $\Omega \times J_{k+1}$ as follows

$S_p\left(\omega ,u\right)=\{\begin{array}{ll}\left(\sigma \left(\omega \right),g^{-1}\left(u\right)\right)\hfill & \text{for}{\omega }_0=1\text{and}\hfill \\ \left(\sigma \left(\omega \right),{\left(2u-g\left(u\right)\right)}^{-1}\right)\hfill & \text{for}{\omega }_0=0.\hfill \end{array}$

Here $g\left(u\right)$ and $2u-g\left(u\right)$ are the increasing self-homeomorphisms of $J_{k+1}$ . The skew product as above preserves the measure ${\mu }_p\times {\mu }_F$ on $\Omega \times J_{k+1}$ . Here $J_{k+1}={\Phi }_{n-1}\left(I_{k+1}\right)$ for $k=1,\cdots ,2n-1$ . The formula ${\left(2u-g\left(u\right)\right)}^{-1}$ for the second homeomorphism is equivalent to $S_p$ invariance of the measure ${\mu }_{\frac{1}{2}}\times \Lambda$ where $\Lambda$ denotes the Lebesgue measure. Section 3 is dedicated to the binomial model for asset prices.

2. The Bernoulli Property of $S_p$

Firstly, we normalize the ${\mu }_F$ measure on $J_{k+1}$ and denote it by ${\overline{\mu }}_F$ .

Definition 2.1. $\left(S_p,{\mu }_p\times {\overline{\mu }}_F\right)$ has the Bernoulli property if

${\overline{S}}_p\left(\overline{\omega },u\right)=\{\begin{array}{ll}\left(\overline{\sigma }\left(\overline{\omega }\right),g^{-1}\left(u\right)\right)\hfill & \text{for}{\overline{\omega }}_0=1\text{and}\hfill \\ \left(\overline{\sigma }\left(\overline{\omega }\right),{\left(2u-g\left(u\right)\right)}^{-1}\right)\hfill & \text{for}{\overline{\omega }}_0=0.\hfill \end{array}$

is a Bernoulli automorphism.

Here $\overline{\omega }\in {\left\{0,1\right\}}^{\mathcal{Z}}$ and $\overline{\sigma }$ is the two-sided $\left(p,1-p\right)$ -Bernoulli shift.

Theorem 2.2. The skew-product transformation $S_p$ has the Bernoulli property.

Proof. By ergodic decomposition of ${\mu }_p\times {\overline{\mu }}_F$ (Kifer, 1986: p. 193, Theorem 1.1) there exists ergodic $S_p$ -invariant measure ${\mu }_p\times {\mu }_G$ such that ${\mu }_G\left(\left\{b_k\right\}\right)={\mu }_G\left(\left\{b_{k+1}\right\}\right)=0$ . Hence ${\mu }_G$ has the dense support property by Lemma 3 (Kowalski, 2009), i.e., if ${\mu }_G\left(A\right)={\mu }_G\left(J_{k+1}\right)$ then $\overline{A}=J_{k+1}$ . Therefore

$M_p\left(S_p\right)=\text{conv}\left\{{\mu }_p\times {\delta }_{\left\{b_k\right\}},{\mu }_p\times {\delta }_{\left\{b_{k+1}\right\}},{\mu }_p\times {\mu }_G\right\}$

by Theorem 1 (Kowalski, 2003). Here $M_p\left(S_p\right)$ denotes the set of $S_p$ -invariant probability measures $\mu$ on $\Omega \times J_{k+1}$ such that the left marginal measure of $\mu$ is ${\mu }_p.$ So we conclude that ${\overline{\mu }}_F={\mu }_G$ as ${\mu }_p\times {\overline{\mu }}_F\in M_p\left(S_p\right)$ and ${\overline{\mu }}_F\left(\left\{b_k\right\}\right)={\overline{\mu }}_F\left(\left\{b_{k+1}\right\}\right)=0$ . Now we are able to use Corollary 5.2 (Kowalski, 2019) to conclude that $\left(S_p,{\mu }_p\times {\overline{\mu }}_F\right)$ has the Bernoulli property. ☐

We define the skew product transformation ${\widehat{S}}_p\left(\omega ,x\right)$ in the space $\Omega \times I_{k+1}$ by putting $H_g$ and $H_{2u-g\left(u\right)}$ instead of $g$ and $2u-g\left(u\right)$ in the definition of $S_p\left(\omega ,u\right)$ . Here

$H_g\left(x\right)={\Phi }_{n-1}^{-1}\left(g\left({\Phi }_{n-1}\left(x\right)\right)\right)\text{for}x\in I_{k+1}.$

This skew product preserves the measure ${\mu }_p\times {\mu }_{{\Phi }_n}$ on $\Omega \times I_{k+1}$ where ${\mu }_{{\Phi }_n}$ has distribution ${\Phi }_n$ . Moreover, its natural extension to automorphism is Bernoulli one. For a detailed description, see (Kowalski, 2022).

3. The Generalized Binomial Model for Asset Prices (Discussion and Conclusions)

A generalized one-asset binomial model, see Bahsoun et al. (2007), Kowalski, (2017), is a random walk on I. Here $x\in I$ represents the security price. Physical phenomena at the quantum level, where randomness rules, can well describe the macro world. Therefore, we can take the skew-product ${\widehat{S}}_p$ as a $\left(n-1\right)$ -asset binomal model. For $n=3$ , see (Kowalski, 2022, Figure 4), we have

$S_d\left(x\right)=h_1\left(x\right)\text{and}{S}_u\left(x\right)=h_0\left(x\right)\text{for}x\in I_4$

where $I_4=\left[0,\frac{\sqrt{2}}{2}\right]$ . This is the price of one share of common stock of a particular corporation. The subscript $u$ of $S_u$ illustrates that the transformation $S_u$ contains a law that moves the price up, and the subscript $d$ of $S_d$ illustrates that $S_d$ contains a law that moves the price down. The Bernoulli property of ${\widehat{S}}_p\left(\omega ,x\right)$ implies the mixing of ${\widehat{S}}_p$ . Hence

$\underset{m\to \infty }{\lim }{\displaystyle \int 1_{\Omega \times A}\left({\widehat{S}}_p{}^m\right)f\text{d}{\mu }_p\times {\mu }_{{\Phi }_n}}={\mu }_{{\Phi }_n}\left(A\right){\displaystyle \int f\text{d}{\mu }_{{\Phi }_n}}$

for every $A\in \mathcal{A}$ and $f\in L_{\infty }\left({\mu }_{{\Phi }_n}\right)$ . Therefore

$\underset{m\to \infty }{\lim }{\mu }_p\left(\left\{\omega :{\widehat{S}}_p^m\left(\omega ,x\right)\in \Omega \times A\right\}\right)={\mu }_{{\Phi }_n}\left(A\right)$

in weak-$L_1\left({\mu }_{{\text{Φ}}_n}\right)$ convergence.

In a similar way, we consider the second financial asset, price-adjacent to the previous one. Here we take

$I_5=\left[\frac{\sqrt{2}}{2}.\frac{\sqrt{6}}{2}\right]$ and

$S_d\left(x\right)=h_0\left(x\right)\text{and}{S}_u\left(x\right)=h_1\left(x\right)\text{for}x\in I_5.$

We can turn the interval $\left[0,\frac{\sqrt{6}}{2}\right]$ into the interval $\left[0,1\right]$ by linearly changing the variables.

For example, the above may be illustrated by the continuous quotations exchange of shares of a company listed on the Warsaw Stock Exchange on May 7, 2024. This is Atende S.A. Here for $n=20$ there are two adjacent zeros of the Hermite polynomials $H_n$ and $H_{n-1}$ such that $x_1\approx 2.79$ , $x_2\approx 3,16$ . If we ignore prices at which the turnover is less than 100, the firm’s price distribution strictly increases as does ${\Phi }_{20}$ . Moreover, $p\approx 0.5$ , so it is a state of equilibrium.

Conflicts of Interest

The author declares no conflicts of interest regarding the publication of this paper.

References

[1]	Bahsoun, W., Góra, P., Mayoral, S., & Morales, M. (2007). Random Dynamics and Finance: Constructing Implied Binomial Trees from a Predetermined Stationary Density. Applied Stochastic Models in Business and Industry, 23, 181-212. [Google Scholar] [CrossRef]
[2]	Kifer, Y. (1986). Ergodic Theory of Random Transformations. In Progress in Probability (Vol. 10). Birkhäuser.
[3]	Kowalski, Z. S. (2003). Invariant Measures for Smooth Extensions of Bernoulli Shifts. Bulletin of the Polish Academy of Sciences. Mathematics, 51, 261-267.
[4]	Kowalski, Z. S. (2009). Iterations of the Frobenius-Perron Operator for Parabolic Random Maps. Fundamenta Mathematicae, 202, 241-250. [Google Scholar] [CrossRef]
[5]	Kowalski, Z. S. (2017). Stability of Smooth Extensions of Bernoulli Shifts. Applicationes Mathematicae, 44, 85-104. [Google Scholar] [CrossRef]
[6]	Kowalski, Z. S. (2019). Bernoulli Property of Smooth Extensions of Bernoulli Shifts. Applicationes Mathematicae, 46, 275-282. [Google Scholar] [CrossRef]
[7]	Kowalski, Z. S. (2022). Random Walk as a Model of Motion in a Quantum Harmonic Oscillator. Applicationes Mathematicae, 49, 185-195. [Google Scholar] [CrossRef]