Improving Sensor Network Coverage Using Modiifed PSO Algorithm

In recent years, the rapid development of wireless sensor network technology make it hasreceived wide attention. Among them, the efficiency of the wholly working system is relatedto coverage quality of WSN. There are a lot of reference standards to evaluate the coveragequality of WSN, such as network coverage ratio, network connectivity, network energyconsumption, communication delay ratio, and so on. In this paper, the network coverage ratiois main reference standards of performance of WSN. And to optimize WSN coverage, as themain purpose is improvement of the network coverage ratio.Particle swarm optimization (PSO) algorithm is one method of swarm intelligencealgorithm, which proposed by scientists’ experiments on observing birds’ foraging behaviorand stimulating the information interaction mechanism during flock foraging process. PSOalgorithm has been widely used because of its advantage of the simple formula and easyunderstanding. On the other side, due to the defects of premature convergence and low searchaccuracy, a large number of works have carried on the improvement of PSO.WSN is generally composed of a large number of sensor nodes, so the coverageoptimization of WSN is a multiple objective optimization problem. In PSO algorithm, eachparticle carries some information, which corresponding to potential solutions of WSNcoverage optimization. And each particle with many dimensions can be correlated to deployedsensor nodes. In the process of particle evolution, poor particles can improve their solutions(deployment locations of sensor nodes) by learning from excellent particles. The algorithmwill eventually get the optimal solution after several iterations. PSO algorithm with thecharacteristics of multi particle dimensions and the strong information interaction ability issuitable for multi-objective dynamic optimization problems suck like the coverageoptimization of WSN.Although, the standard PSO algorithm has certain ability of information interaction, italso has the problem that premature convergence and low performance of optimization hasoften occurred. Thus, the optimization performance in solving the coverage problem of WSNis not well. Adaptive PSO algorithm and VFPSO algorithm make up the shortcomings ofstandard PSO algorithm to a certain extent, but the premature convergence in coverageoptimization of WSN is still restricting coverage ratio improvement. Aiming at the prematureconvergence problem, this paper improves the adaptive PSO algorithm and the VFPSOalgorithm: the comparison between all particles history average optimal value of the previousgeneration and the current generation’s which makes the evolution degree comprehensive isproposed in adaptive PSO algorithm; the dimension selection which solves the prematurealgorithm is proposed in VFPSO algorithm. Similar to the PSO algorithm, BBO algorithm is aclass of multiobjective optimization algorithm which has strong ability of informationinteraction. Its enhanced algorithm (VF-BBO algorithm) has better performance on sensornetwork coverage improvement. So VF-BBO algorithm will be made as the comparisonalgorithm of two kinds of modified PSO algorithm. Through the simulation comparisonbetween adaptive PSO algorithm, VFPSO algorithm and their modified algorithms, two kinds of improved PSO algorithm solve the premature convergence problem better. And, Simulationresults show that two kinds of modified algorithms improve the WSN coverage ratio.