core Estimated learning time: 9 h

PE2-2.2 Probability Theory & Random Processes in Simulation

You can apply Bayes' theorem and work with random variables and distributions, generate random numbers and histograms, and simulate random processes to obtain power spectra of signals and noise.

The point of this unit is that a distribution you can sample is a distribution you can experiment on — simulate ten thousand trials and the law of large numbers shows you the answer before you finish the algebra. Generating a specified distribution from a uniform source is the enabling trick and worth knowing by the inverse-transform method. Histograms are estimates, not truths: bin choice changes what you see, which is a good early lesson in trusting your own plots carefully.

Work through these

  • Probability and Bayes' theorem

    The foundations, and the theorem that lets you reason backwards from evidence. It is used directly in the detection topic later.

  • Random variables; clinical trials as an example

    Random variables introduced through a worked example from outside engineering. The example is chosen so the mathematics is not obscured by the application.

  • Random numbers and random distributions

    Where computer-generated randomness comes from and what distribution it has. Knowing that it is deterministic underneath matters for reproducibility.

  • Histograms and their bin sensitivity

    Looking at a distribution by counting, and the fact that the picture changes with the bin width. It is a good early lesson in how a visualisation can mislead.

  • Functions of random variables

    What happens to a distribution when you transform the variable. It is needed by the generation methods in the next item.

  • Generating random distributions

    How to produce samples from a distribution you want, given samples from one you have. The inverse transform method is the one to understand properly.

  • Laws of large numbers

    Why averages settle down, which is the theoretical backing for every simulation in this course. It also tells you how many runs you need.

  • Random processes and their properties

    Randomness that evolves in time, and the properties that make it tractable. Stationarity and ergodicity are the two assumptions everything later relies on.

  • Power spectra; signals and noise

    Describing a random signal by its frequency content, and the noise model the whole subject uses. It connects this topic to the communication half of the course.

Sign in to keep your progress.

Free resources

Links last checked 30 Aug 2026.

Stuck here?

Ask a mentor. A real person answers, and they can see exactly which topic you're on. Usually within a couple of working days.

Checking your session…

Topics shown in module order.