Further, we assume handover statistics are derived for the
purpose of determination of Pi,j probability of handover from
the serving celli to the target cellj according to [7] or [16]. As
the number of handovers between two cells can be asymmetric,
direction of handover is distinguished for selection of the cells
included in the NCL. It means, the probability of handover
from celli to cellj can be different to the probability of
handover from cellj to celli. This way, the original NCL list Nn
is reduced by consideration of the handover statistics to NH in a
way that NH⊆ Nn. Out of all cells in Nn, only the cells with
Pi,j > Pt are kept in NH. The probability threshold, Pt, is set to
eliminate all unlikely handover target cells from the NCL to
reduce its size.