The amount of scanned cells (including the serving one) is
presented in Fig. 5. As can be expected, the number of
performed scans increases with density of HeNBs for most of
the competitive algorithms. Contrary to these algorithms, our
proposed dynamic threshold scheme is able to cope with
increasing density of HeNBs more efficiently and increase in
the NCL size due to the density of HeNBs is only negligible.
The proposal reaches the lowest amount of scans out of all
schemes disregarding HeNBs’ density. Even if no HeNBs are
deployed, the dynamic threshold reduces amount of scans by
50 % comparing to other approaches. Only hierarchical
algorithm is able to perform similar as our proposal. For higher
densities of HeNBs, the reduction in the amount of scans with
respect to other algorithms is still significant since our proposal
scans only up to 3.1 cells while the second best approach, the
handover history, scans at least 7.5 cells. Thus, the gain
introduced by the dynamic threshold is more than 60% for all
densities of HeNBs. The reason that the handover history
algorithm is getting stable for higher densities is due to the fact
that the UEs are served by the HeNB most of the time for dense HeNB deployment. Thus, a large number of HeNBs
scanned if the UEs are served by the eNB is suppressed in the
results. The number of scanned cells if a UE is attached to a
HeNB is limited to 8 due to the deployment of streets in our
scenario.