Abstract: The artificial hummingbird algorithm is prone to fall into local minima and convergence stagnation when solving high-dimensional complex optimization problems. Therefore，an improved artificial hummingbird algorithm（IAHA）was proposed to optimize wireless sensor network deployment. Firstly，the tangent function transformation of the distance between individual hummingbirds and the optimal hummingbird was performed，and a new foraging strategy was proposed with the optimal hummingbird position as the base and the transformed distance as the flight scale. Secondly，the optimal hummingbird information was perturbed using the Cauchy distribution during the migratory foraging phase，and the perturbation result was assigned to the worst hummingbird，replacing the random assignment method of the original algorithm. Numerical experiments on 12 benchmark functions demonstrated that the IAHA outperformed the six comparison algorithms significantly in terms of finding the best performance. Simulation experiments of wireless sensor network deployment optimization were conducted on four monitoring areas，and the results showed that the average coverage rate obtained by IAHA was higher than that of the comparison algorithm，and the sensors were evenly distributed，which was suitable for solving the wireless sensor network deployment optimization problem.

Key words: artificial hummingbird algorithm, wireless sensor network, Cauchy distribution, tangent transformation distance, coverage rate