1203.4585 (Zhang Jiang et al.)

Ancilla models for quantum operations: For what unitaries does the
ancilla state have to be physical?    [PDF]

Zhang Jiang, Marco Piani, Carlton M. Caves

Any evolution described by a completely positive trace-preserving linear map can be imagined as arising from the interaction of the evolving system with an initially uncorrelated ancilla. The interaction is given by a joint unitary operator, acting on the system and the ancilla. Here we study the properties such a unitary operator must have in order to force the choice of a physical- that is, positive-state for the ancilla if the end result is to be a physical-that is, completely positive-evolution of the system.

View original: http://arxiv.org/abs/1203.4585