【摘要】 求解连续优化问题的蚁群算法的思路为:①根据问题的性质估计下一最优解的范围,并估计出各变量的取值范围;②在变量区域内剖分网格,每一个空间的网格点对应于一个状态,人工蚂蚁在各个空间网格点之间移动,根据各网格点的目标函数值,留下不同的信息量,以此影响下一批人工蚂蚁的移动方向;③循环一段时间后,目标函数值小的网格点信息量较大,据此找出信息量大的空间网格点,并逐步缩小变量范围,在信息量大的空间网格点附近进行人工蚁群移动;④重复前述过程,直到网格的间距小于预先给定的精度,算法终止。本文从蚁群算法的原理出发,将连续域的蚁群算法应用于非线性AVO反演中,并且针对蚁群算法应用过程中容易出现的停滞和扩散问题,对信息素的数量进行了限制。反演结果与理论模型基本吻合,说明了算法的有效性和可靠性,可以应用于其他的地球物理反问题。更多还原

【Abstract】 The idea of the ant colony algorithm which is used to solve continuous optimization problem can be expressed as below: (1)Based on the property of the problem,to estimate the range of the optimum solution and variables’ span,(2)griding the variable region,each grid corresponds to a state,artificial ants moves in the grids,and based on target function value the ants leave different information which will affects moving direction of the next batch of ants,(3)after the iteration running for some time,there will be more information for the grids which have smaller target function values,as a result the grids with more information can be found,the variable region is shrunk,and artificial ants move around the grids with more information,(4)repeating the step (1) to (3),until grid distance is less than pre-set accuracy,then the iteration stopped. Based on principle of the ant colony algorithm,the ant colony algorithm was applied in non-linear AVO inversion,regarding the problems which were often encountered in the application of the ant colony algorithm,such as stagnation and diffusion,the amount of the pheromone was limited. The studies in this paper show that the inversion results are consistent with theoretical model,effectiveness and reliability of the algorithm were proven,it can also be used in other geophysical reverse problems.更多还原