文档介绍:IEEE TRANSACTIONS ON MUNICATIONS, VOL. 2, NO. 2, MARCH 2003 277
A Mixed Neural-ic Algorithm for the Broadcast
Scheduling Problem
Sancho Salcedo-Sanz, Student Member, IEEE, Carlos Bousoño-Calzón, Member, IEEE, and
Aníbal R. Figueiras-Vidal, Senior Member, IEEE
Abstract—The broadcast scheduling problem (BSP) arises in termines the packet average delay. Secondly, for a fixed frame
frame design for packet works (PRNs). The frame struc- length, the channel utilization is determined by the number of
ture determines the munication parameters: communi- simultaneous transmissions of noninterferring stations. We will
cation delay and throughput. The BSP is binatorial opti-
mization problem which is known to be NP-hard. To solve it, we refer to this number as frame throughput or, simply, throughput.
propose an algorithm with two main steps which naturally arise Therefore, the BSP can be stated as to find the minimum length
from the problem structure: the first one tackles the hardest con- and maximum throughput for a PRN.
traints and the second one carries out the throughput optimization. The BSP is binatorial optimization problem known to
This bines a Hopfield work for the con- be NP-hard [2], [3]. Algorithmic solutions have been proposed
straints satisfaction and a ic algorithm for achieving a max-
imal throughput. The algorithm performance pared with based on differe