The Monte Carlo Solver
The Monte Carlo solver samples points randomly with uniform distribution inside a box specified by the user. This solver is slow for finding accurate values of a minimizer of the objective function; however, it is useful for gathering statistical data of design variations by analyzing the range of values the objective function takes. As compared to the other optimization algorithms implemented in COMSOL Multiphysics, it does not get stuck in local minima. It always explores the whole search space specified by the parameter bounds.
The generation of random numbers in the Monte Carlo solver is controlled by the value of the Random seed. If the check box is cleared, the random number generator is initialized by a number based on the current system time. In this case, two runs produce in general different sets of parameters during operation. If a seed is given, the parameter selection is random during the operation of the solver but produces the same sequence of numbers from one run of the optimization solver to the next.