Finding Aggregation Tree with Genetic Algorithm for Network Correlated Data Gathering

The critical issue in designing correlated data networks like Wireless Sensor Networks is to minimize the total cost of data transmission in the network, and decrease the amount of data flow. The problem of finding optimal aggregation tree for correlated data gathering in single sink network is considered as an NP-Complete problem and hence heuristic methods are usually applied to solve it [1]. In this paper, we apply genetic algorithm (GA) to solve the problem. In our method, we improve the performance of genetic search by selecting proper initial population. This initial population is determined in two ways, by using Primepsilas algorithm, and shortest path tree. The main issue is to regard nodes individually, and predict their behavior in global optimum. We use Node-Based genetic algorithm for building the communication tree to improve the encoding and decoding process, locality and heritability versus typical GAs.