Page 40 - 2024-Vol20-Issue2
P. 40

36 |                                                               Jumaa, Mohamad, Allawy & Mohammed

dispensed among ? classes. Now, suppose ni are the nodes in        (once for a Taguchi experiments) i.e., each Taguchi experi-
class i where i = 1 to N. The value of N is assumed to be 4 for    ment is an ad-hoc session for testing the relay node with a
easier simulation; so, the simulation nodes are (n1, n2, n3, n4).  different circumstance.
All nodes have an energy constraint (Ei), and a lifetime ex-       Algorithm (1)
pectation (Li). Each node has a constraint of average power        Step1: let nodei ? Classi where i=1, 2, 3. . . k
which is ?i = Ei/Li. It’s assumed that ?1 > ?2 > ?3 > ?4.          Step2: Let k=4
Since the system is operated in discrete time so, in every slot,   Step3: Let Energy and Lifetime of all nodes are Ei and Li
any node of the N nodes (except n1) could be picked as a           respectively.
source with the same probability where n1 is considered as a       Step4: Average power of nodei is ?i = Ei/Li.
relay node for testing and analyze its behaviour in different      Step5: Let ?1 >?2 > ?3 > ?4
networking circumstances using Taguchi DoE. Each source            Step6: Let n1 is the relay node.
needs several relays to reach the destination and the relays       Step7: Let Bn is number of relayed requested packets made by
max number is (M). For the sake of simplicity, at least one re-    node n.
lay in each session has been assumed. The source demands the       Step8: Let An the relay requested packets number that ac-
relay node to deliver its traffic toward the destination and the   cepted by node n.
relay node could accept or deny the request. The relay node        Step9: Let Dn the relay requests number that made to node n.
transmits its ruling to the source by sending either a negative    Step10: Let Cn the relay requests number that accepted to
or a positive acknowledgment. The traffic session blocked on       node n.
negative acknowledgment. Positive and negative acknowledg-         Step11: Fn = An/Bn , and ?n = Cn/Dn
ments may be sent depending on power constraints. Because          Step12: Suppose e is a tiny +ve number which presents the
the node tested using Taguchi DoE under different conditions
of power constrains. 0.0005 assumed as a constant value of          grade of generosity and tn
energy needed for sending a packet.
For a relay node n1(n), denote by Bn the number of relayed         Step13: If ?n >tn or Fn < ?n - e Reject , Else Accept
requested packets made by node n, and by An the number of
relays requested packets accepted by node n. Dn the relay          VI. TAGUCHI METHOD BASED GTFT NODE
requests number that has been made toward node n, and by                         BEHAVIOUR OPTIMIZATION
Cn the relay requests number that has been accepted to node
n. It’s defined in (7) [1]:                                        The proposed system demonstrates the Taguchi method for
                                                                   studying the effect on GTFT node behaviour; there are four
Fn    =  An  and  ?n  =  Cn           (7)                          factors (parameters). Then use the Taguchi method to infer
         Bn              Dn                                        which factor will influence the node’s cooperative behaviour
                                                                   depending on throughput, ?, and ?; these are the measure-
Watch that ?n is the ratio of the relay requests number by n       ments calculated from the Taguchi method. Where:
which had been accepted, to the requests number that had
                                                                        • Throughput is the number of forwarded packets per
been made by n; thus, Fn is one of the throughput indications             time slot.
which experienced by n.
                                                                        • ? is the (amount of the requested packets that relayed
In the proposed GTFT algorithm, each node preserves an expe-              and created by the node) per (requested packets number
                                                                          that had been accepted by a node from the relays).
rience record for its behaviour depending on the two variables
                                                                        • ? is the (amount of relay request that has been accepted
Fn and ?n. Therefore, each node only keeps information                    to the node) per (number of relay request that has been
for each session kind and does not keep individual logs of                made to the node).

its experience with each node inside the network. The relay        The analysis is based on dependent factors. These factors are:

node always makes decisions based on its values of Fn and               • Power constraints of the node.
?n. Assume that the n relay node receives a request to be
relayed. Suppose that e is a small number representing the              • Several received packets by a specific node that need to
                                                                          be forwarded to the next node.
degree of generosity and that tn is the probable probability of
accepting a request as far as a node in class i fetch a request.        • Power consumption per a packet

GTFT algorithm is as below:                                             • Grade of generosity.

   If ?n > tn or Fn < ?n - e Reject
   Else Accept

The simulation details are summarized in algorithm 1. It
should be mentioned that the algorithm is repeated 9 times
   35   36   37   38   39   40   41   42   43   44   45