基于改进模拟退火算法的横波速度求取

Shear-wave velocity prediction based on revised and simulated annealing algorithm

 王璞 吴国忱 李伟

 中国石油大学（华东） 地球科学与技术学院，山东 青岛 266580

 College of Geosciences and Technology, China University of Petroleum, Qingdao 266580, China

 模拟退火算法能够较好地实现全局最优化求解。基于非均匀变异思想，文中给出了一个新的模拟退火算法扰动模型，该模型具有一定的灵活性，通过选取合适的扰动模型参数，可以达到减少迭代次数、提高收敛速度的目的。在模拟退火算法退温前，增加最优解附近小范围内寻优，以减少对扰动模型的依赖，通过改进，模拟退火算法得到了优化。数值实例分析表明，改进后的模拟退火算法比常规模拟退火算法在迭代次数和扰动次数上有一定的优势。采用该算法，利用DEM模型进行横波速度求取后，与测井横波比较，误差较小，从而验证了新模拟退火算法的实用性和有效性。

 Simulated annealing algorithm is a good way to get the optimal solution. Based on the idea of non-uniform mutation, a new simulated annealing perturbation model is given in this paper. The perturbation model has a certain flexibility in its application. By choosing an appropriate parameter, the iterative number is reduced and the convergence speed is improved. In addition, before the temperature is decreased in the simulated annealing algorithm, the dependence on perturbation model can be reduced by adding a small search scope near the optimal solution. The results show that the revised method is better than traditional method from iterative and perturbing number. A rock physics model is constructed with DEM model to estimate shear-wave velocity under constraint of P-wave velocity. Compared with logging data, the computed shear-wave velocities have high accuracy.

10.6056/dkyqt201503013