Skorokhod Embedding by Randomized Hitting Times
For a general strong Markov process \(X\), the Skorokhod embedding problem consists in finding a (possibly randomized) stopping time \(T\) such that \(X_T\) has a prescribed distribution \(\mu\). As was first shown by \textit{H. Rost} [Z. Wahrscheinlichkeitstheor. Verw. Geb. 15, 19--28 (1970; Zbl 0192.53501)] such a stopping time exists if and only if the potential of \(\mu\) is dominated by that of the initial distribution of \(X\). The present paper shows that the stopping time \(T\) can be chosen to be a `` randomized hitting time. This means that there exists a monotone family of sets \((B(r),\;0\leq r\leq 1)\) such that \(T\) coincides with the hitting time of \(B(R)\) by \(X\), where \(R\) is a random variable uniformly distributed over \([0,1]\), independent of \(X\). The proof depends on tools of the general theory of Markov processes. [For the entire collection see Zbl 0716.00012.]
- Shrinkage estimators, Skorokhod's problem and stochastic integration by parts
- Two explicit Skorokhod embeddings for simple symmetric random walk
- Ordered Skorokhod stopping for a sequence of measures
- Stopping with expectation constraints: 3 points suffice
- On ordered stopping times of a Markov process
- Jensen measures and harmonic measures
- Skorokhod embeddings for two-sided Markov chains
- Extending Chacon-Walsh: minimality and generalised starting distributions
- Embedding of random vectors into continuous martingales
- Stopping distributions for right processes
- Convexity properties of harmonic measures
- The Skorokhod embedding problem and its offspring
This page was built for publication: Skorokhod Embedding by Randomized Hitting Times
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3972793)