Summary: | 碩士 === 義守大學 === 資訊管理學系碩士班 === 97 === In the area of wireless sensor network, Ad Hoc network is a structure similar to the wireless sensor network. A dynamic network topology is an important property of Ad Hoc Network. Sensor nodes can rapidly adapt to the changing network topology. Due to the fact that sensor nodes are moveable, laying sensor nodes is an important fact to obtain an efficient routing. Efficient routing can avoid much redundant transmission when network is exchanging data and thus achieve the purpose of power saving. The construction of Connected Dominating Set (CDS) is a kind of common and frequently used method up to date. Wireless sensor network often use CDS as virtual backbone to get higher efficiency when it transmits. In general, the CDS with smaller size would perform better. For example, with less communication load and simpler link management, etc, so the goal of this study is to find a Minimum CDS (MCDS) in the network topology.
To find out the MCDS efficiently, in previous researches, the common method is to find out the Maximum Independent Set of all sensor nodes in the topology, and then connect them. In this thesis, we used the Genetic Algorithm of artificial intelligence to find out the MCDS. The mating and mutation mechanisms in GA fit the frequent changing characteristics of the network topology. By simulation, comparing, and analyzing, it will put into the spotlight the high efficiency of the simultaneous search of GA. Therefore, in this thesis, we use GA to solve the optimal problem about CDS size.
|