Monte Carlo simulations are used to assess the errors inherent in complex systems and are of particular use when a simple algebraic solution or single error analysis is impossible. The simplest example of a Monte Carlo simulation is the derivation of the probability of a "heads" from a toss of a coin by computationally tossing a coin many times. The success of the simulation relies on how closely the simulation matches reality, the resolution of the simulation and the number of times the simulation is run. Probably the most famous Monte Carlo simulation is that which derives the primordial element abundances from big-bang nucleosynthesis based on the simulation of photon-particle interactions during the recombination era and cooling phase. The power of the method is in the decision as to what is and isn't simulated and the statistical treatment of the results. If time allows I'll also summarise artificial Neural networks, their methodology, application and pros and cons.
Back to course summary.
1998 Apr 23