![](https://csdnimg.cn/release/download_crawler_static/15826576/bg5.jpg)
phabetic Huffman Tree can obviously provide better performance. Xu et al.
[25] considered a novel parameterized index which usually had a linear dis-
tributed structure. In addition, the above schemes are based on off-line to
reduce the access time and energy consumption. Saxena and Pinottti [19] de-
signed a schedule scheme from the perspective of on-line. In the case of no
any a-priori knowledge of either the set of broadcast data items or the prob-
abilities of the requested data items, the data items are divided into multiple
channels in a balanced and a push-pull broadcast schedule is applied in per
channel. In application aspect, Sun et al. [22] proposed a novel two-tier air in-
dexing method for on-demand XML data broadcast. It can reduce index size
through eliminating the redundancy information. Chung et al. [3] adopted
inverted list indexing method to wireless data broadcast for full-text searches.
Based on the latter focus, Lu et al. [16] proposed a heuristic algorithm.
It can employ the algebraic method to download a large number of data
items. In addition, the optimal solution can be obtained in O(2
k
(nt)
O(1)
)
time through proposed AFPT algorithm, where k is the number of the re-
quested data items, n is the number of broadcast channels, and t is the least
access time. Gao at al. [5] developed several efficient data retrieval pro-
tocols for clients to download requested data items. A new problem called
Minimum Constraint Data Retrieval Problem (MCDR) is introduced, and is
proved to NP-hard through using V C ≤
p
MCDR, where V C is the deci-
sion problem of Vertex Cover. In [17], Lu et al. also studied the data retrieval
problem from the point of mobile clients. The largest number of the requested
data items can be downloaded in given time such that the number of switch-
ing channel is minimum (called LNDR) and prove that LNDR is NP-hard
through using 3CN F reduction. Hurson et al. [10] largely reduced energy
consumption through eliminating the maximum number of conflicts, retriev-
ing the maximum number of data elements and minimizing the number of
channel switches. However, this scheme can not always attain the optimal
switch such that the energy consumption can be significantly increased as
the number of the requested data items increases. As a result, a heuristic
scheme is naturally proposed for reducing the number of channel switches in
this paper. Yet, these schemes are mostly concentrated at mobile devices on
one antenna. With the development of wireless networks, the first scheme is
adopted to directly apply in mobile devices on multiple antennae [20].
5