Paper 5: Stochastic Processes and Geometric Flows: From Random Walks to Brownian Motion and Quantum Probability

 

Author: Zhang Suhang

Affiliation: Luoyang, Henan

 

Abstract

 

This paper extends the probability-geometry isomorphism framework from static distributions to stochastic processes, random fields and quantum probability. The main conclusions are demonstrated as follows:

 

1. Random Walks: Path distributions of finite-step random walks correspond to sets of piecewise geodesics in geometric spaces, and step-size distributions determine the length distribution of line segments.

2. Brownian Motion: In the continuous limit, Wiener measure acts as the geometric volume measure on path spaces, with its potential function (action functional) expressed as \frac{1}{2}\int_0^T \|\dot{\gamma}(t)\|^2 dt, which corresponds to the energy functional on Riemannian manifolds. Brownian motion paths are geodesics weighted by this action quantity.

3. Martingales: Square-integrable martingales correspond to minimal surfaces or energy-minimizing mappings in geometry, and the Doob–Meyer decomposition is interpreted as the harmonic decomposition of surfaces.

4. Stochastic Differential Equations: The Fokker–Planck equation is transformed into geometric flows, with the heat equation serving as a special case of Ricci flow. Itô formula is generalized to the chain rule on manifolds.

5. Quantum Probability: Born rule, interference and entanglement can be embedded into projective geometry and complex Hilbert manifolds, while density matrices are regarded as noncommutative geometric potentials.

 

This study achieves complete unification covering dynamics, infinite-dimensional systems, classical and quantum domains, concluding the construction of the probability-geometry isomorphism system.

 

Keywords

Stochastic process; Geometric flow; Brownian motion; Wiener measure; Martingale and minimal surface; Quantum probability; Projective geometry

 

 

 

1 Introduction

 

Papers 1 to 4 have established the isomorphism between static probability and geometry. Each distribution corresponds to a curved surface; marginalization, conditioning and independence match projection, slicing and direct product respectively, with equivalent axiomatic systems. Nevertheless, the real world is dominated by dynamic evolution, such as stock prices, thermal particle motion and information updating, which are described by stochastic processes. Meanwhile, microscopic quantum phenomena break the scope of classical probability, and geometric perspectives offer innovative interpretation approaches.

 

This paper aims to elevate the probability-geometry unification framework to dynamic and infinite-dimensional levels. The core objectives are summarized below:

 

- Paths of stochastic processes are regarded as curves in geometric spaces;

- Probability distributions of paths correspond to geometric volume measures on curve spaces determined by action functionals;

- Properties of martingales reflect minimality of surfaces;

- Noncommutativity in quantum probability is incorporated into complex geometry and projective geometry.

 

Ultimately, Brownian motion in classical probability, path integrals in quantum probability, geodesics and harmonic mappings in differential geometry converge into one unified geometric structure.

 

2 Random Walks: Piecewise Geodesics

 

2.1 Geometric Realization of Discrete-time Random Walks

 

Let S_0 = 0, S_n = \sum_{i=1}^n X_i, where X_i are independent and identically distributed random variables taking values in \mathbb{R}^d with distribution \mu admitting density p(x). A path of length n forms a sequence of points (0, S_1, S_2, \dots, S_n). Geometrically, connecting these points sequentially yields a piecewise linear curve.

 

When the step-size distribution \mu is realized as a geometric contour in accordance with previous studies, the direction and length of each line segment follow this contour. Consequently, the path space of random walks consists of all piecewise geodesics, and the weight of each path equals the product of densities of individual steps.

 

2.2 Geometric Potential of Step-size Distributions

 

Based on the definition of potential function h(x) = -\log p(x) with respect to Lebesgue measure, the probability density of a specific path \gamma = (x_0=0, x_1, \dots, x_n) satisfies:

 

\prod_{i=1}^n p(x_i - x_{i-1}) = \exp\left( -\sum_{i=1}^n h(x_i - x_{i-1}) \right)

 

Geometrically, the term S(\gamma) = \sum_i h(\Delta x_i) represents the action quantity of the path.

 

2.3 Continuous Limit and Preliminaries for Brownian Motion

 

Provided finite step variance and infinite step number n\to\infty with time step \Delta t = 1/n under proper scaling, random walks converge to Brownian motion. Geometrically, piecewise linear curves evolve into smooth curves, and the discrete action sum converges to the integral form \int_0^T L(\dot{\gamma}(t)) dt, where L denotes Lagrangian. For Gaussian step distribution satisfying h(x) = \|x\|^2/(2\sigma^2) + C, the action functional reduces to the energy functional \frac{1}{2\sigma^2}\int_0^T \|\dot{\gamma}\|^2 dt.

 

3 Brownian Motion and Geometry of Wiener Measure

 

3.1 Wiener Measure as Volume Measure on Path Spaces

 

Denote C_0([0,T], \mathbb{R}^d) as the space of continuous paths starting from the origin. Wiener measure W characterizes the probability measure associated with Brownian motion. Its formal density formula is written as:

 

dW(\gamma) \propto \exp\left( -\frac{1}{2}\int_0^T \|\dot{\gamma}(t)\|^2 dt \right) \mathcal{D}\gamma

 

where \mathcal{D}\gamma stands for uniform volume element on path spaces.

 

Geometrically, the path space is treated as an infinite-dimensional Riemannian manifold equipped with metric \langle \delta\gamma, \delta\gamma \rangle = \int_0^T \|\delta\dot{\gamma}(t)\|^2 dt. Wiener measure serves as weighted volume measure on this manifold, with potential function identical to energy functional E(\gamma) = \frac{1}{2}\int_0^T \|\dot{\gamma}\|^2 dt. Brownian motion is essentially Gibbs distribution defined on path spaces, perfectly consistent with the fundamental framework proposed in prior papers, merely extending sample spaces from finite-dimensional Euclidean spaces to infinite-dimensional path manifolds.

 

3.2 Geometric Properties of Brownian Motion

 

- Diffusion and Heat Equation: Transition density of Brownian motion obeys heat equation \partial_t p = \frac{1}{2}\Delta p. Solutions correspond to geometric flows governed by curvature on Riemannian manifolds. Heat kernels in Euclidean spaces coincide with Gaussian distributions represented by parabolic surfaces.

- Most Probable Path: Paths maximizing Wiener measure density under fixed endpoints are energy-minimizing geodesics, consistent with large deviation theory.

- Brownian Bridge: Brownian motion with constrained endpoints corresponds to Gaussian fluctuations around minimal energy surfaces.

 3.3 Logical Coherence with Previous Papers

 

As stated in Paper 4, standardized sums of independent variables converge to normal distributions with parabolic geometric profiles. Brownian motion can be interpreted as infinite-dimensional generalization of central limit theorem, where random walk paths converge and form Gaussian-type measure determined by energy functionals.

 

4 Martingales and Minimal Surfaces

 

4.1 Definition and Geometric Analogy of Martingales

 

A stochastic process M_t is a martingale if \mathbb{E}[M_t | \mathcal{F}_s] = M_s holds for all s<t. Harmonic functions satisfy the mean value property u(x) = \int_{\partial B(x,r)} u(y) d\sigma(y), sharing inherent similarity with martingale characteristics. Harmonic functions can be regarded as deterministic martingales derived from expectations of Brownian motion.

 

A martingale can be mapped onto a random surface, with process values interpreted as surface height along random trajectories.

 

4.2 Doob–Meyer Decomposition

 

The Doob–Meyer theorem claims that any square-integrable martingale admits unique decomposition M_t = M_0 + \text{continuous martingale} + \text{jump component}. Geometrically, continuous martingale parts correspond to harmonic mappings, while increasing processes reflect surface energy. Martingale conditions are equivalent to energy-critical harmonic mappings, transforming martingale theory into the study of stochastic harmonic mappings.

 

4.3 Analogy with Minimal Surfaces

 

Minimal surfaces feature zero mean curvature and locally minimized area. Martingales represent the flattest trajectories minimizing stochastic energy. Brownian motion concentrates near geodesic paths with minimal energy, and high-dimensional martingales are generalized as stochastic minimal surfaces.

 

Conditioning operations correspond to slicing and normalization in finite-dimensional geometry; conditional distributions of incomplete trajectories are determined by harmonic measures, matching Dirichlet problems of minimal surfaces.

5 Stochastic Differential Equations and Geometric Flows

 

5.1 Transformation from SDE to Geometric Flows

 

Consider the stochastic differential equation:

dX_t = b(X_t) dt + \sigma(X_t) dW_t

Density evolution follows the Fokker–Planck forward Kolmogorov equation:

\partial_t p = -\nabla \cdot (b p) + \frac{1}{2} \nabla^2 : (\sigma\sigma^T p)

 

Treating density p as volume density on manifolds, the equation describes geometric flow evolution. Setting b=0 and \sigma = \sqrt{2} yields standard heat equation \partial_t p = \Delta p, which drives smoothing evolution of initial distributions and converges to uniform or Gaussian states.

 

5.2 Connection with Ricci Flow

 

Ricci flow \partial_t g = -2 \mathrm{Ric}(g) describes metric evolution driven by curvature diffusion. Heat equation couples closely with Ricci flow, and random walk and Brownian motion methods construct valid solutions for Ricci flow equations. The unified probability-geometry framework supplies stochastic analysis tools for geometric flow research, and vice versa.

 

5.3 Geometric Formulation of Itô Formula

 

The fundamental Itô formula reads:

df(X_t) = f'(X_t) dX_t + \frac{1}{2} f''(X_t) d[X]_t

On Riemannian manifolds, it is rewritten as:

df(X_t) = \nabla f(X_t) \cdot dX_t + \frac{1}{2} \mathrm{Hess} f (X_t)(dX_t, dX_t)

consistent with manifold Taylor expansion, further consolidating the intrinsic connection between probability and geometry.

 

6 Quantum Probability: Projective Geometry and Complex Hilbert Manifolds

 

6.1 Basic Framework of Quantum Probability

 

Quantum states are represented by unit vectors or density matrices in Hilbert spaces, observables correspond to self-adjoint operators, and measurement probabilities comply with Born rule: P(\lambda) = \langle \psi, P_\lambda \psi \rangle, where P_\lambda denotes spectral projection operator. Quantum probability forms a noncommutative extension of classical probability theory.

 

6.2 Geometric Realization via Complex Projective Spaces

 

Pure quantum states correspond to points on complex projective spaces \mathbb{P}(\mathcal{H}) equipped with Fubini–Study metric. Geometric interpretations are listed as follows:

 

- Quantum state \psi ↔ Point on projective manifold

​

- Observable quantity A ↔ Real-valued geometric function \langle A \rangle_\psi = \langle \psi, A\psi \rangle

​

- Measurement probability ↔ Geometric measure on submanifolds generated by orthogonal projection

 

Mixed states described by density matrices reside on probability simplices over projective spaces, extending classical probability into noncommutative domains.

 

6.3 Fusion with Classical Probability-Geometry Isomorphism

 

Classical probability distributions correspond to Riemannian surfaces with volume element p(x)dx. Density matrices serve as noncommutative geometric potentials in quantum systems, whose eigenvalues form probability distributions and eigenvectors define orthogonal spatial directions. Quantum state collapse is geometrically interpreted as orthogonal subspace projection, analogous to conditional probability slicing operations in finite-dimensional cases. Quantum probability naturally generalizes classical isomorphism within complex and noncommutative geometric frameworks.

7 Unified Overview: From Finite Dimension to Infinite Dimension, Classical to Quantum

 

Hierarchical correspondence between probabilistic objects and geometric realizations:

 

表格

Level Probabilistic Object Geometric Realization 

Static Finite Dimension Density distribution   Curved surface with volume element   

Static Infinite Dimension Random field   Weighted volume measure on field spaces 

Dynamic Finite Dimension Stochastic process   Wiener-type measure on path spaces 

Dynamic Infinite Dimension Stochastic partial differential equation Curvature-driven geometric flow 

Noncommutative Quantum System Density matrix   Measure on complex noncommutative manifolds 

 

All cases follow the unified paradigm: probabilistic objects equal geometric volume measures weighted by exponential potential functionals.

 

8 Conclusion and Prospect

 

This paper complements the final component of the probability-geometry unification system, expanding static isomorphism into dynamic stochastic evolution and quantum probability domains. Key achievements include:

 

- Random walk paths are equivalent to piecewise geometric geodesics;

​

- Wiener measure of Brownian motion acts as intrinsic volume measure on infinite-dimensional path spaces;

​

- Martingale theory corresponds to harmonic mappings and minimal surface geometry;

​

- Fokker–Planck equations characterize manifold geometric flows;

​

- Classical probability-geometry isomorphism is generalized to quantum projective and noncommutative geometry.

 

Papers 1 to 5 jointly establish a complete theoretical system: probability theory is essentially geometric measure theory, probability spaces are geometric manifolds equipped with volume measures, and stochastic processes correspond to natural geometric flows. Gaussian bell curves mark the initial sign of isomorphism, axiomatic equivalence builds core foundations, and stochastic & quantum extensions verify universal applicability.

 

Subsequent research will conduct academic comparison with Grothendieck geometry and information geometry, and explore practical applications in machine learning, statistical inference and quantum computation.

 

References

 

[1] Zhang S H. Fundamental Paradigm of Probability-Geometry Isomorphism, 2026

[2] Zhang S H. Geometric Realization of One-dimensional Probability Distributions, 2026

[3] Zhang S H. Geometric Embedding of Multivariate Random Variables, 2026

[4] Zhang S H. Geometric Reconstruction of Probability Axiom System, 2026

[5] Wiener N. Generalized harmonic analysis. Acta Mathematica, 1930

[6] Itô K. On stochastic differential equations. Memoirs of the American Mathematical Society, 1951

[7] Doob J L. Stochastic Processes. Wiley, 1953

[8] Nelson E. Dynamical Theories of Brownian Motion. Princeton University Press, 1967

[9] Malliavin P. Stochastic calculus of variations and hypoelliptic operators, 1978

[10] von Neumann J. Mathematical Foundations of Quantum Mechanics, 1932

[11] Connes A. Noncommutative Geometry, 1994