Tensions in Practice: A small control surface in tension with independent movement¶
Two state coordinates · one bounded update
A toy device starts at (0, 0). One actuator adds the same input u to both coordinates, so they can move only together. A second actuator can add v to the second coordinate alone. This opens new targets, but bounded inputs still rule some out in one update. More control changes the reachable set; it does not remove every constraint.
Keep the control surface small
Use one actuator when coupled movement is all the task needs.
Move coordinates independently
Reach selected off-diagonal targets by adding a separately controlled input.
Why these aims pull against each other
The shared actuator ties two coordinates together. Adding a new direction of intervention costs another actuator and control path, while its magnitude limit still bounds reach.
Choose an arrangement to see what changes and what remains difficult.
Finite illustrative comparisons. Text states carry the meaning; color is not a measured score or universal preference.
What this choice protects
What it costs
When it fits
Compare the arrangements
Shared input
Starting from (0, 0), one update gives (u, u), where u is −1, 0 or 1.
| Input u | Input v | Reachable? | |
|---|---|---|---|
| Target (1, 1) | 1 | Absent | Yes |
| Target (1, 0) | No choice | Absent | No |
| Target (0, 1) | No choice | Absent | No |
| Target (−1, 1) | No choice | Absent | No |
- What it protects
- One actuator and one input path suffice for the required diagonal targets.
- What it costs
- Off-diagonal targets cannot be reached with this input structure.
- When it fits
- The task requires coupled movement only and the extra actuator’s engineering and failure burden is not justified.
Illustration note: The finite setting and values are editorial assumptions, not measured effects or recommended operating settings. “No choice” means no admissible u reaches the target in the stated update. It is not an optimization failure.
Extra input
Starting from (0, 0), one update gives (u, u+v). Each input is independently restricted to −1, 0 or 1.
| Input u | Input v | Reachable? | |
|---|---|---|---|
| Target (1, 1) | 1 | 0 | Yes |
| Target (1, 0) | 1 | −1 | Yes |
| Target (0, 1) | 0 | 1 | Yes |
| Target (−1, 1) | −1 | Needs 2 | No |
- What it protects
- The targets (1, 0) and (0, 1) become reachable through explicit admissible input pairs.
- What it costs
- Another actuator and control path require engineering and add failure or misuse possibilities; target (−1, 1) still needs an unavailable v=2 in one update.
- When it fits
- Independent movement is needed and the added hardware/control responsibility is worth maintaining.
Illustration note: The finite setting and values are editorial assumptions, not measured effects or recommended operating settings. This table is one-step bounded reachability, not a claim of global controllability for a continuous physical plant.
What this illustration does—and does not—establish
Controllability: Controllability cost versus value — actuator proliferation and safety risk supplies the extra-lever cost and Controllability: Controllability constraints and the design of reachable regions the input-bound distinction. The algebraic device and selected targets are editorial.
- No unforced drift, disturbance or measurement error is included.
- The one-update horizon is fixed; repeated updates would create a different reachable set.
- Reachability does not establish that a trajectory is optimal, robust or safe in a real device.
Source entries
Controllability
Controllability: Controllability cost versus value — actuator proliferation and safety risk supplies the conflict examined here.
Controllability cost versus value — actuator proliferation and safety risk
Expanding controllability (adding actuators, deploying feature flags, increasing admin API surface) costs engineering effort and increases risk (each new control lever can be misused or can fail in unanticipated ways).
Structural Tensions
Real systems have bounded inputs (thrusters with limited thrust, software deployments that take time, policy instruments with political limits); these constraints define what's reachable within a budget.