The planner must choose a trajectory, not only the next node.
If the chosen configuration must remain admissible over the whole dwell interval, the problem changes from a one-step choice to selecting an evolution policy.
From configuration selection to policy selectionExpand section
From configuration selection to policy selection
As long as the decision is evaluated only in the current state, the problem looks one-step. Once the minimum dwell time τd is introduced, that is no longer enough: the chosen configuration must remain admissible throughout the whole dwell interval.
J(π) → max
A special case, not the general lawExpand section
A special case, not the general law
The one-step criterion is correct only when the ordering of configurations by survivability does not change over the dwell interval and no admissible configuration loses admissibility before the next decision epoch.
This condition is often violated under sudden failures and jump-like epistemic degradation.
What trajectory planning buysExpand section
What trajectory planning buys
On the reference model, trajectory planning does not improve average survivability under predictable degradation, but in sudden-failure scenarios it reduces the time spent with an inadmissible resource from 3.55% to 1.90%, with only a 0.12% reduction in integral survivability.