Information Technology Reference
In-Depth Information
heuristic know in what state is SQRP. Is to say, if is necessary to add more TTL, because the
amount of resources found are few and decreases the lifetime.
Tables Storage: Short and Long learning,
applied in selecting the next neighbor.
Tables Storage: Short and Long learning,
applied in selecting the next neighbor.
τ
τ
D
D
N
N
H
H
Hyper-Heuristic
Ant Colony Algorithm
HH_AdaNAS
Hyper-Heuristic
Ant Colony Algorithm
HH_AdaNAS
τhh
τhh
η
η
HH_TTL. Tables storage long learning,
applied in selecting the next low-level
heuristic.
HH_TTL. Tables storage long learning,
applied in selecting the next low-level
heuristic.
Fig. 9. Storage structures of HH_AdaNAS.
The pheromone table τ is divided into n two-dimensional tables, one corresponding to each
node i of the network. These tables contain only entries for a node fixed i , therefore, its
dimensions are at most | L |×|Γ ( i )|, where L is the dictionary, which defines the keywords
allowed for consultation and Γ ( i ) is the set of neighboring nodes of i . Each in turn contains a
two-dimensional table | m |×| h |, where m is the states visibility set of the problem and h is
the available heuristics set. The pheromone table is also called learning structure long.
The visibility state table η expresses the weight of the relation between SQRP-states and
TTL-heuristics and was inspired by the deterministic survival rule designed by Rivera
(Rivera G. 2009). Table η is formed by the combination of | m |×| h |, where a visibility state
m i is identified mainly by α, which depends on the node selected by AdaNAS to route the
query SQRP. The variable α in Equation 6 contributes to ensure that the node selected by
HH_AdaNAS, in the future, not decreases the performance of the algorithm. A TTL-
heuristic is intelligently selected according with the past performance given by its
pheromone value, and its visibility value, given by an expert. The Figure 10 shows the
visibility state table used in this work.
h 1
h 2
h 3
h 4
m 1
1
0.75
0.5
0.25
m 2
0.75
1
0.5
0.5
m 3
0.5
0.5
1
0.75
m 4
0.25
0.5
0.75
1
Fig. 10. Visibility state table
=(
H/ D) / Z
(6)
i,j,l
i,j,l
x
Search WWH ::




Custom Search