A Genetic Algorithm Based and Subtree Restricted Routing Algorithm in Wireless Sensor Networks

The GA (genetic algorithms) was applied to find routing trees with minimal maximal sensor load thus prolong network lifetime in wireless sensor networks. However, the existing algorithm limits search space of GA to avoid generating cycles by dividing sensors into different layers and stipulating that a sensor can only select a sensor in its lower layer as its new parent. In this paper, we have found that how cycles are generated during GA operations, and proposed a solution based on subtree to avoid generating cycles. The solution only requires a sensor not to change its parent to a sensor in the subtree with it as the root. And layers are no longer considered. As a result, GA has more search space and GA operations can be performed more freely. Thus more optimal routing trees can be computed and network lifetime can be extended. The experiment shows that the proposed algorithm prolongs network lifetime notably.