Monte Carlo methods. Vol. 1: basics
Monte Carlo methods. Vol. 1: basics
| Hi-index | 0.00 |
The use of M independent computational processors by distributing random samples among them decreases the cost of the Monte Carlo method by M times, as the final summation and averaging of the results are practically inessential. This approach is especially effective when using the 'double-randomization' method for solving the problems with random parameters. When M is large, the necessary amount of random numbers is also very large, and it is especially expedient to use the combined random-pseudorandom secuence. For global estimating a solution in the metric C by simulation of series of trajectories from different points, it is reasonable to use the same random numbers for each point. The fact decreases the necessary amount of random numbers.