Suppose a single server processes entities from two groups each consisting of three sources. The switching component between the entity sources and the server determines which entities proceed to the server whenever it is available. The switching component uses a distribution that is skewed toward entities from the first group. Within each group, the switching component uses a round-robin approach.
The example below (open modelmodel) shows how to implement this design using three Input Switch blocks. The first two Input Switch blocks have their Switching criterion parameter set to Round robin to represent the processing of entities within each group of entity sources. The last Input Switch block uses a random signal with a skewed probability distribution to choose between the two groups. The Signal Latch block causes the random number generator to draw a new random number after each departure from the last Input Switch block.
For tracking purposes, the model assigns an attribute to each entity based on its source. The attribute values are 1, 2, and 3 for entities in the first group and -1, -2, and -3 for entities in the second group. You can see from the plot below that negative values occur less frequently than positive values, reflecting the skewed probability distribution. You can also see that the positive values reflect a round-robin approach among servers in the top group, while negative values reflect a round-robin approach among servers in the bottom group.