Padding a 0-work-tape NTM with a dummy work tape #
NTM.pad0 turns an NTM 0 into an NTM 1 deciding the same language in the
same time bound: the added work tape is never used. The dummy head steps right
off ▷ on its first move (respecting the δ_right_of_start discipline) and
then parks at cell 1, writing □ over the □ already there, so the dummy
tape's cells never change.
This closes the k = 0 case of the single-tape reduction
(NTM.exists_singleTape_decidesInTime): the simulation machinery requires at least
one work tape, while a padded machine is already single-work-tape.
Pad a 0-work-tape machine with one never-used work tape. The dummy tape's
action: step right off ▷, otherwise stay put and write □ (over the □
already under the head).
Equations
- One or more equations did not get rendered due to their size.
Instances For
Trace correspondence. The padded machine's run tracks the original's component-wise; the dummy work tape keeps its initial cells with the head parked at cell 0 or 1.
Padding preserves the all-paths halting bound.
Padding preserves timed acceptance, in both directions.
Padding preserves deciding. A 0-work-tape decider yields a 1-work-tape decider for the same language in the same time bound.