Monte Carlo from Scratch - Poor Little Kid 穷孩子

From the lab to the computer¶

Monte Carlo simulation of radiation transport is a scientific instrument by its own right. Cultivating the skill and the art is no less than gaining mastery in instrumentation. Monte Carlo is never a plug-and-play software tool. Sadly, Monte Carlo is often abused as a black box. Throughout the years I have also been ‘offered’ co-authorship of to-be papers which were to become publishable if I ran Monte Carlo for them. No, Monte Carlo cannot make an otherwise unpublishable work publishable.

A piece of study begins with a study or experimental design, considering the choice of samples and the control of variables. Letting all variables lose is not going to lead us to anything conclusive.

Embarking on a Monte Carlo project, I would think of the experimental setup the similar way as I would do when setting up a physical experiment in the radiation laboratory. The difference is that now:

we need not worry about ALARA or ALARP, radiation exposure or radiation protection; we can do cleaner experiments, we shall be able to control controlled variables much better; we can do experiments which would otherwise be

unethical e.g. shooting test beams into human bodies;
impractical e.g. knock down shielding walls and rebuild new ones iteratively as part of the design process;
impossible e.g.
- detecting particles directly and perfectly without perturbing the radiation field and without - requiring the particles to interact and deposit energy;
- finding out about the particle’s ancestry at the time we detect the particle; we call this technique ‘latching’.