Download presentation
Presentation is loading. Please wait.
Published byAshley Dustin Hampton Modified over 9 years ago
1
Two Connected Dominating Set Algorithms for Wireless Sensor Networks Overview Najla Al-Nabhan* ♦ Bowu Zhang** ♦ Mznah Al-Rodhaan* ♦ Abdullah Al-Dhelaan* The Proposed Algorithms Connected Dominating Set (CDS) CDS is known to be an efficient strategy to control network topology, reduce overhead, and extend network lifetime. CDS finds a minimum size subset of nodes that can collaboratively form a virtual backbone. Definition For a given connected graph (network) G = (V, E), a dominating set (DS) is a subset V' of V, where for each vertex (node) u of V, u is either in V' or at least one neighbor vertex of u is in V'. A DS is called a CDS if the sub-graph induced by the vertices in the DS is connected. In this work, we present the design of two novel algorithms for CDS construction in WSNs. Our algorithms are intended to minimize CDS size. The first algorithm has a performance factor of 5 from the optimal solution, which outperforms the best-published results (S-MIS algorithm) that has a performance factor = 5.8+ ln 4. The second algorithm is an improved version of the first algorithm. Approach-I 2.The newly colored black node (u ) dominates its 1-hop yellow, orange, and white nodes by coloring them gray. Furthermore, u marks its 2-hop white/orange nodes into yellow; and marks its 3-hops white nodes into orange. Approach-I Approach-II Approach-II also has 4 phases. Phase-1, 3, and 4 are similar to their corresponding phases in Approach-I. In Phase-2, we define the coverage factor of a yellow/gray node x as the number of its yellow neighbors. A gray/yellow node x that has the highest coverage factor is marked red. Then, x dominates its 1-hop yellow neighbors by coloring them gray. In Phase-3: If a gray node u was marked red in Phase-2 of this approach, u is already connected to an S 1 node and we do not need to introduce any connectors for u. Simulation Results N [ 36,400 ], the transmission range of each node R [200, 800], and the considered deployment schemes are: the uniform random and the partial random deployment models. Results show that the sizes of CDSs generated by Approach-I and Approach-II are smaller than S-MIS as the network size increases. Approach-II always outperforms both Approach-I and S-MIS in uniform random and grid-based deployments, and for small and large-scale networks. To implement distributed versions of the proposed algorithms. To introduce more performance improvements. Future Work Approach-I consists of 4 main phases to construct a CDS: Phase-1: S1 Construction given an arbitrary rooted spanning tree T, we define the tree level of a node u as the number of hops in T between u itself and i, where i is the root of T. All nodes are initially undominated and colored white. The root node (i) initiates S 1 construction by coloring itself black. Then: all white nodes that are 1-hop from i are colored gray; all white nodes that are 2-hop from i are colored yellow; all white nodes that are 3-hop from i are colored orange. Next, the algorithm repeats the following steps until no orange/white nodes left in the graph. 1.Selects an orange node u to color it black. The selected orange node satisfies the following two conditions: i) its level is the lowest (closes to the root) among all orange nodes, and ii) it has the maximum number of 3-hop black neighbors. The selected node u is colored black. We model a WSN using a unit disk graph (UDG). We simplify the CDS construction by first finding a special independent set S 1 satisfying the following condition: the hop-distance between any two complementary subsets S’ and S” of S 1 is exactly three. Second, we find a small set of nodes S 2 to dominate the multiple disconnected components resulted from constructing S 1 in the first step. Then, nodes in S 2 are connected with nodes in S 1 in order to form the final CDS by adding more connecting nodes. The above described technique is performed in two different ways in Approach-I and Approach-II. For illustration purpose, we employ a coloring scheme to differentiate node states during the construction process. Dominators (S 1 ): black, dominatee: gray, (S 1 ): red, connectors: blue. Other colors are temporary colors. *{nalnabhan,rodhaan,dhelaan}@ksu.edu.sa, **{bowuzh}@gwmail.gwu.edu Phase-2: Covering Disconnected Regions (CDC) We optimally compute a minimum dominating set for each connected yellow component. All the dominators computed from this phase are colored red and they form the set S 2. Phase-3: Connecting S 2 Nodes to S 1 Nodes. Phase-4: Connecting S 1 Nodes all Together. An exemplary graph G of 40 nodes after S 1 construction Performance Analysis Given any Minimum CDS (MCDS) of a unit-disk graph G, we show that: Approach-I produces a CDS with a size bounded by 5opt, where opt is the size of the MCDS. The proof of this theorem is provided in the paper. For the simulation, we focused on CDS size as the main and most important performance measure. We generated a total of N nodes in a fixed 1000*1000 2D square. Random network of 70 nodes Node density [28, 346] N [36, 225], R=400
Similar presentations
© 2024 SlidePlayer.com. Inc.
All rights reserved.