In some cases rather than using randomly selected values you may want to methodically search out all possible parameter combinations in an N dimensional grid. For example you might want to see what happens when x takes on the values 1, 2, 3, 4 and 5 and y takes on the values 3,5 and 7. There are 5 x 3 = 15 combinations of values that can be checked.
If you specify a Latin Grid search Vensim will perform this methodical search. It will ignore the number of simulations specified and instead do enough simulations to look at every combination of parameters that is possible from the different RANDOM distributions or VECTORs specified.
Clearly, for even a modest number of parameters this can entail a large number of simulations indeed. The total sample size is the product of the sizes of each dimension. For RANDOM elements, this is controlled by the Number of Grid Entries control parameter. For VECTOR elements, it is determined by (max-min)/interval. Thus sampling combinations of 5 grid entries for 3 parameters will require 125 simulations.