Study on Hybrid Intelligent Algorithm with Solving Pre-stack AVO Elastic Parameter Inversion Problem

The process of pre-stack AVO(Amplitude Variation with Offset) elastic parameter inversion is the process of optimization, which can be solved by using genetic algorithm. However, when solving the problem, the traditional genetic algorithm converges speedily and is easily trapped into the problems like local optimum, etc., which leads to unsatisfactory inversion effect. To solve the above problems, this paper introduces simulated annealing algorithm into genetic algorithm and improves the genetic algorithm in the aspects of population initialization, selection strategy and generic manipulation. A hybrid intelligent algorithm is proposed in this paper which is more suitable for solving pre-stack AVO elastic parameter inversion problem. The experimental results show that the proposed hybrid intelligent algorithm in this paper can fully exert the global searching capability of genetic algorithm and prevent the algorithm from being trapped into local optimum by using simulated annealing algorithm. The Gardner relation is utilized to initialize population in order to make initialization of density correspond to the practical terrain conditions better and obviously improve the inversion accuracy.