Show that the claim of Theorem 9.35 continues to hold if X...

Show that the claim of Theorem $9.35$ continues to hold if $X$ is only a submartingale but if $\varphi$ is in addition assumed to be monotone increasing. Give an example that shows that the monotonicity of $\varphi$ is essential. (Compare Corollary 9.34.)

Key Concepts

Submartingales

Submartingales are stochastic processes that satisfy the condition that the conditional expectation, given past information, is at least as large as the present value. This property, which generalizes the idea of martingales, is fundamental in the theory of stochastic processes as it models systems with an inherent upward bias in expectation. Understanding submartingales is crucial when applying transformations that need to preserve such ordering in conditional expectations.

Monotone Increasing Functions

A monotone increasing function is one that preserves the order of its inputs; that is, if one value is less than or equal to another, then its image under the function is also less than or equal to the image of the other value. In the context of probabilistic transformations, applying a monotone increasing function to a process like a submartingale is important because it helps to maintain the inequality relationships that define the submartingale property.

Jensen's Inequality and Convex Functions

Jensen's inequality is a fundamental result which states that for a convex function, the function applied to an expectation is less than or equal to the expectation of the function applied to the variable. When combined with monotonicity conditions, convex functions can be used to transform martingales or submartingales in a controlled manner, ensuring that the key properties of the original process are preserved under the transformation.

Role of Monotonicity in Transformations

The requirement of monotonicity in the transformation function is essential when working with submartingales because it guarantees that the order of the process is not reversed by the transformation. Without monotonicity, the inequality inherent to the definition of a submartingale can be lost, which means that the transformed process may no longer satisfy the submartingale property. An example illustrating this typically shows that if a non-monotone function is applied, the conditional expectation might not respect the necessary inequality, thus emphasizing the critical role of monotonicity.

Transcript

Please give Ace some feedback