**2.1 The routing protocol low energy adaptive clustering hierarchy (LEACH)**

Low Energy Adaptive Clustering Hierarchy LEACH is considered to be the first cluster-based hierarchical routing protocol proposed by Heinzelman et al. As one of the most popular hierarchical routing algorithms for wireless sensor networks [4]. The LEACH protocol assumes equal residual energy from the sensor when starting the network. The life of the network is then divided into towers by a choice of CH. However, each cycle consists of two phases: an initialization phase and a transmission phase (**Figure 2** below displays LEACH cluster formation).

**Figure 2.** *LEACH cluster formation [5].*

#### **2.2 Clustering with K-means**

We distinguish several types of clustering techniques: The most used are Partitioning, Hierarchical, Density and Grid based. These algorithms try to decompose the set of nodes into a number of disjoint clusters. The problem is how to select the Cluster Head (CH) and how to manage the clusters. There are many methods of partitioning clustering. The most famous ones are: Fuzzy C-Means, K-means, K-medoids and Partitioning Around Medoids (PAM). The authors [6] surveyed and summarized all the partitioned clustering protocols

**181**

**Table 1.**

*K-Means Efficient Energy Routing Protocol for Maximizing Vitality of WSNs*

in WSNs. The next table (**Table 1**) examined a list of routing protocols in wireless sensor networks (WSNs) that uses partitioning clustering with the

learning algorithms that solve clustering problems, developed by J.B.Mac Queen in 1967. In this method we assign a point in each group where the center of this group is closest (centroids). The center is the average of all the points of the group and its coordinates are the arithmetic average of each dimension, together with all the points of the group, which means that each cluster is represented by its center of gravity. **Figure 3** below shows the flowchart of K-means

**Ref Characteristics Intra-cluster inter–cluster CH choice Compared** 

[8] KPSO/KGA One -hop One -hop K-means/

[11] EBRP One -hop One -hop K-means++

[12] Algorithm genetic One -hop One -hop K-means/

*The different routing protocol uses k-means in WSNs [3].*

One -hop One -hop Nœud près

One -hop One -hop Midpoint

multi -hop One -hop Euclidian

The K-MEANS algorithm is presented as one of the simplest non-supervised

**with**

Clustering traditionnel

du centre de gravité

Distance euclidienne

algorithm mathematical format

distance residual energy

Euclidian Distance

Euclidian Distance

**Advantages**

k-means • Reduces

k-means • More

/ • Better per-

k-means • Better than

GA • Optimal

• Evolving. • Optimal number of clusters. • Increased network life.

energy consumption. • Optimal number of clusters. • Increased network life.

> efficient and correct than k-Means for large clusters.

formance.

k-Means for large clusters.

solution.

*DOI: http://dx.doi.org/10.5772/intechopen.96567*

"K-MEANS" method.

[7] k-means used to implement an optimal number of

clusters.

[9] Unique node ID refers to the Euclidian distance between nodes and center of gravity.

[10] Combined (K means, Davies Bouldin Index, Gaussian elimination)

algorithm.

#### *K-Means Efficient Energy Routing Protocol for Maximizing Vitality of WSNs DOI: http://dx.doi.org/10.5772/intechopen.96567*

in WSNs. The next table (**Table 1**) examined a list of routing protocols in wireless sensor networks (WSNs) that uses partitioning clustering with the "K-MEANS" method.

The K-MEANS algorithm is presented as one of the simplest non-supervised learning algorithms that solve clustering problems, developed by J.B.Mac Queen in 1967. In this method we assign a point in each group where the center of this group is closest (centroids). The center is the average of all the points of the group and its coordinates are the arithmetic average of each dimension, together with all the points of the group, which means that each cluster is represented by its center of gravity. **Figure 3** below shows the flowchart of K-means algorithm.


#### **Table 1.**

*Computational Optimization Techniques and Applications*

**2.1 The routing protocol low energy adaptive clustering hierarchy (LEACH)**

The LEACH protocol assumes equal residual energy from the sensor when

Low Energy Adaptive Clustering Hierarchy LEACH is considered to be the first cluster-based hierarchical routing protocol proposed by Heinzelman et al. As one of the most popular hierarchical routing algorithms for wireless sensor networks [4].

starting the network. The life of the network is then divided into towers by a choice of CH. However, each cycle consists of two phases: an initialization phase and a transmission phase (**Figure 2** below displays LEACH cluster formation).

We distinguish several types of clustering techniques: The most used are Partitioning, Hierarchical, Density and Grid based. These algorithms try to decompose the set of nodes into a number of disjoint clusters. The problem is how to select the Cluster Head (CH) and how to manage the clusters. There are many methods of partitioning clustering. The most famous ones are: Fuzzy C-Means, K-means, K-medoids and Partitioning Around Medoids (PAM). The authors [6] surveyed and summarized all the partitioned clustering protocols

**2. Related work**

*A wireless sensors network architecture [3].*

**Figure 1.**

**180**

**Figure 2.**

**2.2 Clustering with K-means**

*LEACH cluster formation [5].*

*The different routing protocol uses k-means in WSNs [3].*

*Computational Optimization Techniques and Applications*

**Figure 3.** *Flowcharts of K-means algorithm.*
