Computational Intelligence And Its Application In Wireless Sensor Network Optimization

As science develops and technology progresses, more and more problems have the feature of large-scale, complexity, constraint, nonlinearity and nondeterminacy. Optimizing of all kinds of process in production and science is an urgent problem that needs to be solved, however, traditional optimization methods are difficult to solve these complicated large-scale problems. As a rising technology of optimization, Computational Intelligence provide innovative thought and means to solve these complicated problems. The algorithms based on computational intelligence are easy to understand and realize, especially, most of computational intelligence methods possess inner parallel and distribution, auto-organization, auto-adaptation, which effectively promote the application of computational intelligence in the domain of optimization. Computational intelligence play more and more important roles in improving the production efficiency, reducing consuming energy and saving resource.This paper starts with two methods applied widely of Computational Intelligence:Evolutionary Computing and Swarm Intelligence. It makes the Genetic Algorithms(GA) that is classic in Evolutionary Computing and Ant Colony Optimization(ACO) algorithms that is representative in Swarm Intelligence as its study foundation. It presents theory and characteristic of the two methods simply, then gives some improvements and makes some combinations with other methods to seek the application of intelligent optimization in engineering practice. In application, in view of the feature that wireless sensor networks(WSN) must possess auto-organization, auto-adaptation and robustness, especially, energy of WSN is very limited, this paper fully utilizes the advantages of Computational Intelligence, marries together both the research focuses. It proposes some methods and ideas for applying Computational Intelligence to solve optimization problems of WSN. The major contents of this paper could be summarized as follows:1. The problem of clustering in WSN is described and studied in this paper. Synthetically considering information of position and energy of neighbor nodes in cluster, GA is applied to optimize the selection of cluster head to balance the consuming energy of nodes in cluster, avoid nodes dying early, and prolong the network life-time. 2. This paper depicts coverage problem of WSN, for the feature that this problem is the problem of multiobjective optimization, under the topology control of GAF, it applies GA based on Pareto sorting to solve the problem, then improves this algorithm to maintain population diversity and obtain high-quality, well distributed Pareto solutions. The algorithm it proposes realizes the aim that using the least number of sensor nodes to achieve the best coverage, which is able to save energy of the network, decrease the interference between signals and prolong the network life-time.3. Aiming at routing problem of WSN, after analyzing the features of hierarchical routing algorithms in WSN, this paper synthetically considers the depleted energy during the data transferring and the residual energy of node, proposes a hierarchical routing algorithm based on ACO. Finally, the algorithm improves performance of multi-hop routing of cluster head, and balances the consuming energy of nodes in WSN.