Decision Trace

While a run is animating, Decision Trace puts a small chip on the canvas each time a routing decision fires, saying why the entity went the way it did. The chips appear beside the source node’s output side, fade in, hold for about a second and a half, and fade out.

There is nothing to switch on: running the model produces them.

You will see nothing at the shipped animation speed. The chips render only at a speed of 20x or slower, and at most five are on screen at once, while the speed slider starts at its maximum, well above 20x. Drag the speed slider well down before you expect to see a chip.

Figure 1. A decision chip during an animated run, with the speed slider along the foot.

What a label says

Every chip reads the same way: the reason, then the route it took.

• Percent: the share that sent it, such as 60% and the target it chose. An expression-driven percent shows as a function instead of a number.
• Conditional: the condition that was true, and the target.
• Balancing: which rule picked the target, Shortest Queue, Least WIP, Lowest Utilization or Shortest Cycle.
• Schedule, Available and Alternate: the route taken because of a schedule, because a target was available, or because the alternate was used.

A plain single route, or a default one, produces no chip at all. Only a real decision is annotated, so anything you see on the canvas is a fork the model actually chose between.

Saving and replaying a trace

Two separate things sit on the File menu, and they can be used together.

File ▸ Save trace as copies the session’s trace file, a .pmdtrace, to a location you pick. It stays greyed out until a run has produced a trace to save.

File ▸ Open Simulation Trace opens a saved animation trace and puts the model into replay mode, so you can watch a run back without running it again.