Optimization of particle swarm algorithm and its usage in calculation of hydrogeological parameter

Abstract

In order to improve the accuracy, sensitivity, and convergence, the particle swarm optimization (PSO) algorithm is optimized and applied in the calculation of hydrogeological parameters to improve the accuracy of the hydrogeological model. The sensitivity and accuracy of the algorithm are improved through derivation of the appropriate value of the particles. Based on convergency analysis of this algorithm with the three intelligent optimization algorithms (Sphere function, Rosenbrock function, and Griewank function), the convergence of the algorithm is analyzed through different intelligently optimized algorithms. At the same time, the accuracy is validated through comparison between the optimized and unoptimized particle swarm algorithm as well as the actually observed hydrogeological data, and the sensitivity of the water conductivity coefficient and water storage coefficient under the algorithm is analyzed. The results show that the calculated value of the optimized PSO algorithm is very close to the theoretical value 0 of the intelligent optimization algorithm Sphere function and the theoretical value 1 of the Rosenbrock function as well as the theoretical value 1 of the Griewank function. The results reveal that the calculation results are very close to the theoretical value in the intelligent optimization algorithm test of optimized particle swarm algorithm; the maximum absolute error between the calculated value and the observed value of the optimized particle swarm algorithm is 0.011, and the maximum relative error of which is 8.1%; the maximum absolute error between the calculated value of particle swarm algorithm and the observed value is 0.021, and the maximum relative error reaches 29.1%; the number of iterations of the optimized particle swarm algorithm is 87 on average, while it of the unoptimized particle swarm algorithm is 450, reaching the optimal value. In addition, the optimized PSO algorithm has a standard function of 0.00659, which is significantly smaller than that of the PSO algorithm of 0.00684. When the interference coefficient is −20%~20%, the water conductivity coefficient and water storage coefficient are in negative phase with the drop depth of the aquifer. It suggests that the optimized particle swarm optimization algorithm has high accuracy and convergence, and its sensitivity of water conductivity coefficient and sensitivity of water storage coefficient are both good, which can provide a reliable algorithm basis for the construction of hydrogeological model and the establishment of aquifer parameters.