Swarm intelligence optimization algorithm, as a kind of intelligence computation,which is attravtive for more and more researchers. Particle swarm optimization(PSO)algorithm, which is one of swarm intelligence optimizations, is a hot topic in research. Itis a relatively new stochastic optimization technique that can simulate the swarmforaging behavior of birds flocking. PSO has faster convergence speed, but it is easy tofall into local optimum in dealing optimization problems. Stardard PSO algorithmâ€™sevolution model, topological structure and learning strategies are studied in the paper.According to the existing advantages and dis advantages of PSO, we propose three newimproved particle swarm optimization algorithms, and the first improved algorithm isapplied to the wireless sensor network coverage optimization problem. Specific researchcontents are as follows:(1) In order to overcome the drawbacks of particle swarm optimization (PSO) thateasily falls into local minima and lacks in diversity, this paper proposes PSO algorithmwith adaptive subspace Gaussian learning. The discrete degree of fitness and subspaceGaussian learning are employed to adaptively adjust parameters and search strategies,and help the algorithm to jump out of the likely local optima. Moreover, we proposeneighborhood learning strategy in which the optimal neighborhood particle isintroduced, where the neighborhood of the current particle is dynamically constructedduring the evolution, increases the diversity of population. Finally, our approach isapplied to wireless sensor network coverage optimization problem and obtains apromising performance, it can make a more uniform distribution of wireless sensornodes and higher coverage of the networks.(2) As standard particle swarm optimization (PSO) algorithm has someshortcomings, such as getting trapped in the local minima, converging slowly and lowprecision in the late of evolution. A new improved PSO algorithm basing on Gaussiandisturbance (GDPSO) is proposed in this paper. Gaussian disturbance is putted into inthe personal best positions, which can prevent falling into local minima, and improvethe convergence speed and accuracy.(3) To solve the standard PSO slow convergence in the late of evolution, a new improved opposition-based particle swarm optimization algorithm (IOPSO) is proposedin this paper. Each particle generates an opposite particle to expand the scope of thesearch area, and enhance the global search ability of the algorithm. Moreover, eachparticle must learn form itâ€™s opposite particle to avoid the best particle being trappedinto local optima, since this may cause search stagnation of the whole swarm. Theresults show that our improved algorithm obtains better performance in globalconvergence speed, accuracy and the ability to escape from local optimum. |