To prove the existence of a canonical Brownian motion, one commonly follows the following steps:

Step 1: Define the Sample Space

  • The sample space Ω is typically chosen as a suitable function space, such as the space of continuous functions on a state set S (often denoted as C(S)).
  • The functions in Ω represent the sample paths of the Brownian motion, and each function corresponds to a possible trajectory of the process.

Step 2: Define the Sigma-Algebra

  • The sigma-algebra F on the sample space Ω (e.g., C(S)) is constructed to define measurable sets or events.
  • The sigma-algebra can be the Borel sigma-algebra, which is generated by the open sets in the topology of pointwise convergence or uniform convergence.
  • The sigma-algebra F determines the collection of measurable sets on which probabilities can be assigned.

Step 3: Define the Probability Measure

  • The probability measure P is defined on the measurable space (Ω, F), providing probabilities for the measurable sets or events.
  • To define the probability measure, one often starts by specifying the finite-dimensional distributions of the Brownian motion.
  • The Kolmogorov extension theorem is typically used to extend the finite-dimensional distributions to a unique probability measure on the entire sample space (Ω, F).

Step 4: Verify Properties

  • The constructed probability space (Ω, F, P) must satisfy specific properties to ensure the existence of a canonical Brownian motion.
  • These properties include sample path continuity, independent and stationary increments, and Gaussian distribution properties for the increments.
  • Verification involves using mathematical techniques, such as moment generating functions, characteristic functions, and limit arguments.

Step 5: Construct the Brownian Motion

  • With the probability space (Ω, F, P) established, the canonical Brownian motion B(t) can be defined as a stochastic process on this space.
  • The process B(t) should satisfy the desired properties, such as B(0) = 0, independent and stationary increments, and Gaussian distribution properties for the increments.

The construction of a canonical Brownian motion requires careful attention to the choice of sample space, sigma-algebra, and probability measure. The verification of specific properties ensures that the resulting process satisfies the desired characteristics of a Brownian motion. Advanced mathematical tools, including measure theory and stochastic analysis, are typically employed in the construction and verification process.

I hope this explanation provides a more specific and detailed overview of the process involved in proving the existence of a canonical Brownian motion.


표본공간을 연속함수들의 모임으로 정하고 

시그마대수는 보렐시그마대수 사용하고 

kolmogorov 써서 분포를 정해주는것도 정확한거같은데 성능좋네