Abstract
We study violations of n-particle Bell inequalities (as developed by Mermin and Klyshko) under the assumption that suitable partial transposes of the density operator are positive. If all transposes with respect to a partition of the system into p subsystems are positive, the best upper bound on the violation is In particular, if the partial transposes with respect to all subsystems are positive, the inequalities are satisfied. This is supporting evidence for a recent conjecture by Peres that positivity of partial transposes could be equivalent to the existence of local classical models.
- Received 15 October 1999
DOI:https://doi.org/10.1103/PhysRevA.61.062102
©2000 American Physical Society