**5. Energy aware nearest cell association (EANCA) algorithm**

An energy aware MR placement is accomplished by developing an EANCA algorithm. The conditions for associating the MC with a MR are the residual energy must be *Energy Aware Router Placements Using Fuzzy Differential Evolution DOI: http://dx.doi.org/10.5772/intechopen.83747*

greater than the consumed energy and the power consumed by the both the mesh nodes must be less than the maximum power allocated. Even if any MR has no client association and it is in idle state, still there is some minimum amount of energy consumption. In order to overcome this, the proposed energy aware scheme turns the inactive mesh routers to sleep mode thus minimizing the energy consumption [13]. The energy consumption level of a node at any time of the simulation can be determined by finding the difference between the current energy value and initial energy value.

The methodology is illustrated as follows:


If Dmin = f(short, medium) and.

If Pt = f(medium, high) then MR ∈ active mode.

If TL = f(low, very low) and.

If Dmin = f(long, very long) and.

If Pt = f(low) then MR ∈ sleep mode.

Step 6: Check the QOS constraints

Step 7: If true then

```
Mc(j) ∈ MR(i)
  Er(new) = Er-1
Else if.
 Mc(j) ∉ Mr.(i)
Then.
```
Compute the fitness value.

Step 9: End

The performance of the proposed scheme is analyzed based on three important metrics PDR, throughput and FR.

### **5.1 Performance metrics**

*Throughput*. Network throughput is the average rate of successful message delivery over a communication channel.

*Packet delivery rate*. The ratio of the average number of data packets received by the destination node to the number of data packets transmitted.

*Failure rate (FR)*. In a time slot there is a possibility for a MC not to be assigned to a MR, hence they get disconnected which is referred as connection failure. The network performance is evaluated through a metric failure rate and it is defined as the number of failures to the attempts to make the connection.

### **6. Simulation results and discussion**

The proposed approach is evaluated in NS2 simulator and compared with the existing algorithms. The MRs is deployed in a large terrain area of 1000 m 1000 m

**5. Energy aware nearest cell association (EANCA) algorithm**

*Wireless Mesh Networks - Security, Architectures and Protocols*

**Figure 4.** *Fuzzy rule viewer.*

**Figure 5.** *Surface view plot.*

**142**

An energy aware MR placement is accomplished by developing an EANCA algorithm. The conditions for associating the MC with a MR are the residual energy must be and the clients are distributed normally. The deployment field is equally divided into grid cells with equal area. The simulation period is set as 12 hours i.e. half a day, which is divided into 108 consecutive slots each with time duration of 400 seconds. The failure rate threshold is set as 1. The network performance metrics are evaluated for this simulation model. The simulation is repeated for 200 generations and the FR percentage is calculated after each generation. The input simulation parameters are given in **Tables 3** and **4**.

the given number of clients and it is found that there is a gradual decrease in energy consumption from 80th generation to 120th. From 120th generation the schemes have converged for the optimal result. Whereas SA and the conventional method using Traffic Weight (TW) allotment of gateways show high energy consumption as the number of routers required are high as well as the routers are inactive with no

In order to overcome the premature convergence in DE, fuzzy DE method uses

The FDE approach shows 5.12% lesser energy consumption than TW method, 4.4% lesser energy consumption than SA and 0.75% lesser than DE algorithm. The network performance is evaluated through metrics such as throughput and PDR. Throughput refers how much data can be transferred from one location to

PDR is defined as the ratio between the successfully received packets in the destination to the number of data packets sent from the source node. The comparative results of the conventional methods and evolutionary approaches are tabulated in **Table 5** for 45 clients, 16 routers and one gateway in normal distribution. The observed results show that the FDE approach produces better results compared to the other approaches. The number of failures for the clients to associate

**Data rate = 12 Mbps; no. of clients = 45; no. of MRs = 16 Methodology Throughput (Mbps) PDR (%) Failure rate (%)** Conventional 6 61 27.35 TW 6.8 67 20.2 SA 8.9 71 13.25 DE 11 97 12.12 FDE 11.12 97.34 10.1

the knowledge base fuzzy rules. When CR is high than the scaling factor the convergence rate is faster. The proposed FDE scheme utilizes also the knowledge about the network load, router and client distance which enables the system to

consume less power than other node placement schemes.

*Energy Aware Router Placements Using Fuzzy Differential Evolution*

*DOI: http://dx.doi.org/10.5772/intechopen.83747*

another in a given amount of time.

client association.

**Table 5.**

**Figure 7.**

**145**

*Percentage of failure rate vs. no. of mesh clients.*

*Performance evaluation.*

The energy consumption is evaluated for FDE approach and is compared with the DE, SA and conventional method using only the traffic weight to allot the gateway. From the simulated results shown in **Figure 6**, it is observed that in DE and FDE placement scheme the number of routers required is minimum to cover


### **Table 3.**

*Simulation parameters.*


### **Table 4.**

*Simulation settings-optimization methods.*

**Figure 6.** *Energy consumption of mesh nodes.*

### *Energy Aware Router Placements Using Fuzzy Differential Evolution DOI: http://dx.doi.org/10.5772/intechopen.83747*

the given number of clients and it is found that there is a gradual decrease in energy consumption from 80th generation to 120th. From 120th generation the schemes have converged for the optimal result. Whereas SA and the conventional method using Traffic Weight (TW) allotment of gateways show high energy consumption as the number of routers required are high as well as the routers are inactive with no client association.

In order to overcome the premature convergence in DE, fuzzy DE method uses the knowledge base fuzzy rules. When CR is high than the scaling factor the convergence rate is faster. The proposed FDE scheme utilizes also the knowledge about the network load, router and client distance which enables the system to consume less power than other node placement schemes.

The FDE approach shows 5.12% lesser energy consumption than TW method, 4.4% lesser energy consumption than SA and 0.75% lesser than DE algorithm. The network performance is evaluated through metrics such as throughput and PDR. Throughput refers how much data can be transferred from one location to another in a given amount of time.

PDR is defined as the ratio between the successfully received packets in the destination to the number of data packets sent from the source node. The comparative results of the conventional methods and evolutionary approaches are tabulated in **Table 5** for 45 clients, 16 routers and one gateway in normal distribution.

The observed results show that the FDE approach produces better results compared to the other approaches. The number of failures for the clients to associate


### **Table 5.**

and the clients are distributed normally. The deployment field is equally divided into grid cells with equal area. The simulation period is set as 12 hours i.e. half a day, which is divided into 108 consecutive slots each with time duration of 400 seconds. The failure rate threshold is set as 1. The network performance metrics are evaluated for this simulation model. The simulation is repeated for 200 generations and the FR percentage is calculated after each generation. The input simulation param-

*Wireless Mesh Networks - Security, Architectures and Protocols*

The energy consumption is evaluated for FDE approach and is compared with the DE, SA and conventional method using only the traffic weight to allot the gateway. From the simulated results shown in **Figure 6**, it is observed that in DE and FDE placement scheme the number of routers required is minimum to cover

**Parameters Values** Area size 1000 m 1000 m MAC 802.11 s No. of mesh routers 16 No. of mesh clients 45 Application type CBR Packet size 1024 bytes Transmission power 15 dBm

**Parameters SA DE FDE** Placement of nodes Random Random Random Population size 100 100 100

CR Probabilistic selection 0.5 Fuzzy rule based selection

eters are given in **Tables 3** and **4**.

**Table 3.**

**Table 4.**

**Figure 6.**

**144**

*Energy consumption of mesh nodes.*

*Simulation parameters.*

*Simulation settings-optimization methods.*

*Performance evaluation.*

**Figure 7.** *Percentage of failure rate vs. no. of mesh clients.*

The FDE approach shows 5.12% lesser energy consumption than TW method, 4.4% lesser energy consumption than SA and 0.75% lesser than DE algorithm. The failure rate is also observed to be less than the other schemes. The proposed FDE method has 20% less failure rate than DE and 23.7% less than SA schemes. The conventional method shows a steep increase as the number of mesh clients increases whereas the evolutionary approaches tries to settle down in optimum points. As the number of clients increases FDE and DE approaches converge and show only 11% of FR. Thus the results obtained using FDE shows less energy

*Energy Aware Router Placements Using Fuzzy Differential Evolution*

*DOI: http://dx.doi.org/10.5772/intechopen.83747*

Department of ETCE, Sathyabama Institute of Science and Technology, Chennai,

© 2019 The Author(s). Licensee IntechOpen. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/ by/3.0), which permits unrestricted use, distribution, and reproduction in any medium,

\*Address all correspondence to: merlinsheebu@gmail.com

provided the original work is properly cited.

consumption and failure rate.

**Author details**

G. Merlin Sheeba

Tamil Nadu, India

**147**

**Figure 8.** *Convergence rate.*

with the mesh routers are high for conventional and TW methods compared to the evolutionary schemes. The results show that the FDE approach has 20% less failure rate than DE and 23.7% less than SA schemes. Even if the network size increases with more number of clients the proposed FDE approach is able to show less failure percentage as displayed in **Figure 7**.

The conventional method show a steep increase as the number of mesh clients increases whereas the evolutionary approaches tries to settle down in optimum points. As the number of clients increases FDE and DE approaches converge and show only 11% of FR. The comparison of the three evolutionary approaches with convergence graphs are shown in **Figure 8**. Standard deviation is calculated for each approach. The algorithm is run for 1000 iterations and it is observed that the result of FDE approach converged with 0.001787, which is a very low value from 300th iteration. The results obtained from FDE approach is consistent and has faster convergence speed compared to DE and SA.

### **7. Summary**

To summarize, energy aware placement using FDE approach is proposed to minimize the energy consumption. A transmission cost metric is defined as a function. Three important parameters, the minimum distance between the MRs and MCs, the transmission power of routers and traffic load.

The deployment field is divided into cells of equal area wherein the candidate locations of each MR is positioned. Normal distribution is selected to distribute the clients as it shows 36.6% increase of PDR than SA approach in the previous module.

Usually the DE control parameters are fixed but the FDE scheme uses the CR and S values adaptively to settle for optimum point. A fuzzy inference engine is used to map the input to the output function. The uncertain network parameters are also mapped using the fuzzy inference engine to evaluate the transmission cost.

An energy aware nearest cell association algorithm is proposed to make the MRs to sleep if they are in idle state. If the MRs have no associated clients then the MR is considered to be idle. Any network device in idle state consumes power hence a sleep mechanism is introduced to place energy aware routers.

### *Energy Aware Router Placements Using Fuzzy Differential Evolution DOI: http://dx.doi.org/10.5772/intechopen.83747*

The FDE approach shows 5.12% lesser energy consumption than TW method, 4.4% lesser energy consumption than SA and 0.75% lesser than DE algorithm. The failure rate is also observed to be less than the other schemes. The proposed FDE method has 20% less failure rate than DE and 23.7% less than SA schemes. The conventional method shows a steep increase as the number of mesh clients increases whereas the evolutionary approaches tries to settle down in optimum points. As the number of clients increases FDE and DE approaches converge and show only 11% of FR. Thus the results obtained using FDE shows less energy consumption and failure rate.
