If the are trying to control which of two paths the traveler takes to get to the same destination as in this layout, then the solution is to return a value in the OnArrival trigger of NN2 equal to the edge number of the path you want the traveler to take. That's the nice thing about the OnArrival trigger's of NetworkNodes. If you return a value greater than 0, the traveler will be "nudged" to take the path numbered the same as the value. In this model, the edge number from NN2 to NN3 is 2, and the edge number from NN2 to NN4 is 3; therefore, it works to just use this expression in the OnArrival trigger of NN2:
if(fromedge == 1)
{
return bernoulli(30, 2, 3);
}
The reason this works so easily is because all the travelers have a destination for their travel task of NN5 and of course the Sink4 is their destination for the final unload task. Therefore, no matter which of the two nodes they get nudged to (NN3 or NN4), they can still proceed forward from that point to their final destinationn of NN5.
If NN5 were removed from the network, and NN3 and NN4 were both connected to Sink4 as shown in the picture below, it would require a little different approach. In this case you would have to redirect the traveler to a different travel destination than possibly it was headed to when it left Source1.
Don't be confused between the destination for their travel task, and the destination for their unload task. You might remember from training that there is both regular travel and "offset travel". The offset travel is the travel that occurs from the node to the unload/load station and is associated with the load/unload task, not the travel task.
To change the travel destination, there is a command called redirectnetworktraveler() that you can use. The code in NN2's OnArrival trigger will need to both nudge and redirect the destination and will look something like this:
if(fromedge == 1)
{
int newEdge = bernoulli(30, 2, 3);
treenode newDestination = outobject(current, newEdge);
redirectnetworktraveler(traveler, newDestination);
return newEdge;
}
The last situation you might come across is when you not only need to change the network travel destination, but also the unload destination (for instance you might have a Sink2 and a Sink3 in the above example model). In this case you would be forced to preempt the traveler's active tasksequence it was dispatched with at Source1 with a new tasksequence. Just remember if you need to preempt and dispatch another tasksequence to a traveler on the network, do not do it directly in the OnArrival trigger. Instead, call senddelayedmessage() in 0 seconds, and do your dispatching in the OnMessage trigger!
How to control path of travelers on network route?
I tried to reproduce your last example. However, after preemting the current task sequence with a new one the travel task isn't executed anymore and the TE starts with it offset travel. So the question is, what did I do wrong?