Continuous-time Markov chain

In probability theory, a continuous-time Markov process is a stochastic process { X(t) : t ≥ 0 } that satisfies the Markov property and takes values from a set called the state space; it is the continuous-time version of a Markov chain. The Markov property states that at any times s > t > 0, the conditional probability distribution of the process at time s given the whole history of the process up to and including time t, depends only on the state of the process at time t. In effect, the state of the process at time s is conditionally independent of the history of the process before time t, given the state of the process at time t. In simple terms the process can be thought of as memory-less.

Mathematical definitions

File:FinancialMarkovProcess.png

Directed graph representation of a Markov process describing the state of financial markets (note: numbers are made-up).

A Markov process, like a Markov chain, can be thought of as a directed graph of states of the system. The difference is that, rather than transitioning to a new (possibly the same) state at each time step, the system will remain in the current state for some random (in particular, exponentially distributed) amount of time and then transition to a different state. The process is characterized by "transition rates" $q_{ij}$ between states i and j. Let X(t) be the random variable describing the state of the process at time t, and assume that the process is in a state i at time t. $q_{ij}$ (for $i\neq j$) measures how quickly that $i\rightarrow j$ transition happens. Precisely, after a tiny amount of time h, the probability the state is now j is given by

$\Pr(X(t+h)=j|X(t)=i)=q_{ij}h+o(h),\quad i\neq j,$

where o(h) represents a quantity that goes to zero faster than h goes to zero (see the article on order notation). Hence, over a sufficiently small interval of time, the probability of a particular transition (between different states) is roughly proportional to the duration of that interval. The $q_{ij}$ are called transition rates because if we have a large ensemble of n systems in state i, they will switch over to state j at an average rate of $nq_{ij}$ until n decreases appreciably.

The transition rates $q_{ij}$ are typically given as the ij-th elements of the transition rate matrix Q. As the transition rate matrix contains rates, the rate of departing from one state to arrive at another should be positive, and the rate that the system remains in a state should be negative. The rates for a given state should sum to zero, yielding the diagonal elements to be

$q_{ii}=-\sum _{j\neq i}q_{ij}.$

With this notation, and letting $p_{t}=\Pr(X(t)=j)$, the evolution of a continuous-time Markov process is given by the first-order differential equation

${\frac {\partial }{\partial t}}p_{t}=p_{t}Q$

The probability that no transition happens in some time r is

$\Pr(X(s)=i~\forall ~s\in (t,t+r)\,|\,X(t)=i)=e^{q_{ii}r}.$

That is, the probability distribution of the waiting time until the first transition is an exponential distribution with rate parameter $-q_{ii}$, and continuous-time Markov processes are thus memoryless processes.

A time dependent (time heterogeneous) Markov process is a Markov process as above, but with the q-rate a function of time, denoted q_ij(t).

Embedded Markov chain

One method of finding the stationary probability distribution, π, of an ergodic continuous-time Markov process, Q, is by first finding its embedded Markov chain (EMC). Strictly speaking, the EMC is a regular discrete-time Markov chain, sometimes referred to as a jump process. Each element of the one-step transition probability matrix of the EMC, S, is denoted by s_ij, and represents the conditional probability of transitioning from state i into state j. These conditional probabilities may be found by

$s_{ij}={\begin{cases}{\frac {q_{ij}}{\sum _{k\neq i}q_{ik}}}&{\text{if }}i\neq j\\0&{\text{otherwise}}.\end{cases}}$

From this, S may be written as

$S=-D_{Q}^{-1}Q,\,$

where D_Q = diag{Q} is the diagonal matrix of Q.

To find the stationary probability distribution vector, we must next find $\phi$ such that

$\phi (I-S)=0,\,$

with $\phi$ being a row vector, such that all elements in $\phi$ are greater than 0 and $||\phi ||_{1}$ = 1, and the 0 on the right side also being a row vector of 0's. From this, π may be found as

$\pi ={-\phi D_{Q}^{-1} \over \left\|\phi D_{Q}^{-1}\right\|_{1}}.$

Note that S may be periodic, even if Q is not. Once π is found, it must be normalized to a unit vector.

Another discrete-time process that may be derived from a continuous-time Markov chain is a δ-skeleton—the (discrete-time) Markov chain formed by observing X(t) at intervals of δ units of time. The random variables X(0), X(δ), X(2δ), ... give the sequence of states visited by the δ-skeleton.

Applications

Markov processes are used to describe physical processes where a system evolves in constant time. Sometimes, rather than a single systems, they are applied to an ensemble of identical, independent systems, and the probabilities are used to find how many members of the ensemble are in a given state. A master equation treatment is often used to analyse systems that evolve as Markov processes , with approximations possible for complicated systems .

Chemical reactions

Imagine a large number n of molecules in solution in state A, each of which can undergo a chemical reaction to state B with a certain average rate. Perhaps the molecule is an enzyme, and the states refer to how it is folded. The state of any single enzyme follows a Markov process, and since the molecules are essentially independent of each other, the number of molecules in state A or B at a time is n times the probability a given molecule is in that state.

Queueing theory

As a simple example, imagine a first-come-first-served queue where the jobs have exponential size distribution with average size $\mu$ and a Poisson arrival rate $\lambda$. Then the number of jobs in the queue is a continuous time Markov process on the non-negative integers. The transition rate from one number to the next (meaning a new job arrives within the next instant) is $\lambda$, and the transition rate to the next lowest number (meaning a job completes in the next instant) is $\mu$.

See also

• Markov chain
• Master equation (physics)
• Queueing theory
• Phase-type distribution
• Semi-Markov process
• Variable-order Markov model

References

• William J. Stewart (1994). Introduction to the Numerical Solution of Markov Chains. Princeton University Press. pp. 17–23. ISBN 0-691-03699-3.