Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Let $(X_t, Y_t)$ be a two-dimensional Markov stochastic process that runs on time interval $[t_0, t_f]$. Its infintesimal generator is described by the functions $\mu_X, \mu_Y, \sigma_X, \sigma_Y$. I have the following information about the process:

(1) $\mu_X(x, y, t)$ and $\sigma_X(x, y, t)$ are known. For both these functions, the parameter $y$ has no effect (so we can write them $\mu_X(x, t)$ and $\sigma_X(x, t)$).

(2) $\mu_Y(x, y, t) = x$

(3) $\sigma_Y(x, y, t) = 0$

(4) $X_{t_0} = x_0$, some known constant. $Y_{t_0} = 0$.

My end goal is: looking forward from $t_0$, find a probability density function that reflects my beliefs about the value of $Y_{t_f}$.

I'm only vaguely familiar with stochastic processes, so I would appreciate a description of how to solve this problem or a reference to a gentle introduction to solving these sorts of problems.

Thank you.

share|cite|improve this question
up vote 2 down vote accepted

In other words, $(X_t)_t$ can be any (autonomous) diffusion solving the stochastic differential equation $\mathrm dX_t=\sigma_X(X_t,t)\mathrm dW_t+\mu_X(X_t,t)\mathrm dt$, and $$ Y_t=\int_0^tX_s\mathrm ds. $$ At this level of generality, I am not sure that much more can be said.

Edit: Recall that the transition probabilities $(p_t)_t$ of $(X_t)_t$ solve the forward Kolmogorov (or Fokker-Planck) equation $$ \partial_tp_t+\partial_x(\mu_Xp_t)-\tfrac12\partial^2_{xx}(\sigma^2_Xp_t)=0. $$ Likewise, the transition probabilities $(P_t)_t$ of $(X_t,Y_t)_t$ such that, for every suitable function $u$, $$ \mathbb E(u(X_t,Y_t))=\iint u(x,y)P_t(x,y)\mathrm dx\mathrm dy, $$ solve the forward Kolmogorov (or Fokker-Planck) equation $$ \partial_tP_t+\partial_x(\mu_XP_t)+x\partial_yP_t-\tfrac12\partial^2_{xx}(\sigma^2_XP_t)=0. $$

share|cite|improve this answer
Yeah, time-integration is precisely the motivation behind this post. I got a suggestion over at MO to attempt time-integration by phrasing as a two-dimensional Markov process, but I'm unclear on the procedure for converting between infinitesimal generators and transition functions in multiple dimensions. – GMB Jan 6 '13 at 7:42
Multiple dimensions do not change a iota to the theory. See Edit. (Is this your question?) – Did Jan 6 '13 at 11:25
Yes, it's precisely my question. Thanks. – GMB Jan 7 '13 at 7:49

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.