Static Monte Carlo（3-39）.pdfVIP

1 Static Monte Carlo This chapter introduces fundamental methods and algorithms for simulating samples of random variables and vectors, and provides illustrative examples of these tech- niques in quantitative finance. Section 1.1 introduces the simulation problem and the basic Monte Carlo valuation. Section 1.2 describes several algorithms for im- plementing a simulation scheme. Section 1.3 treats some methods for reducing the variance in Monte Carlo valuations. 1.1 Motivation and Issues Monte Carlo is a beautiful town on the Mediterranean coast near the border between France and Italy. It is known for hosting an important casino. Since gambling has been long considered as the prototype of a repeatable statistical experiment, the term “Monte Carlo” has been borrowed by scientists in order to denote computational techniques designed for the purpose of simulating statistical experiments. A simula- tion algorithm is a sequence of deterministic operations delivering possible outcomes of a statistical experiment. The input usually consists of a probability distribution de- scribing the statistical properties of the experiment and the output is a simulated sam- ple from this distribution. Simulation is performed in a way that reflects probabilities associated with all possible outcomes. As such, it is a valuable device whenever a given experiment cannot be repeated, or it only can be repeated at a high cost. In this case, first a model of the conditions defining the original experiment is established. Then, a simulation is performed on this model and taken as an approximate sampling of the true experiment. This method is referred to as a Monte Carlo simulation. For instance, one may generate scenarios about the future evolution of a financial mar- ket variable by simulating samples of a market model defining certain distributional assumptions. Monte Carlo methods are very easy to implement on any computer system. They can be employed for