Skip to content
Probability II / Continuous-Time Markov Chains and Queueing Theory

Computing Transition Probabilities: Kolmogorov's Forward and Backward Equations

Given jump rates, how do you compute pt(i,j)=P(Xt=j∣X0=i)p_t(i,j)=\mathbb{P}(X_t=j\mid X_0=i) for an arbitrary tt? This note derives the two Kolmogorov differential equations — obtained by splitting [0,t+h][0,t+h] at either end — and shows both are solved by the matrix exponential pt=eQtp_t = e^{Qt}, the direct continuous-time analogue of pmp^m being a matrix power in discrete time.
Learning objectives. By the end of this note you should be able to:
  • state and derive the continuous-time Chapman-Kolmogorov equation;
  • derive Kolmogorov's backward and forward equations from it;
  • state the matrix-exponential solution and verify it for small examples;
  • solve the two-state chain's transition probability in closed form.

In this lesson

  1. The Chapman-Kolmogorov equation
  2. The backward and forward equations
  3. Worked examples

The full lesson, with worked examples and exercise solutions, is for members.

Read the full lesson

New to SelfMadeQuant.com? See membership options or explore the curriculum.