利用双种群遗传算法进行数值试井自动拟合

遗传算法以随机化技术为指导，通过对整个解空间的高效搜索而得到全局最优解，当解空间较大时，常规遗传算法难以同时保证搜索速度和最优解的精度。针对这种缺陷提出了应用双种群遗传算法，即利用具有不同搜索策略的2个种群进行联合搜索，既保证了最优解的精度，又提高了搜索速度，将其应用于数值试井自动拟合解释，比常规遗传算法能节省50％以上的计算时间。

Numerical well testing auto-matching applying double population genetic algorithm

Genetic Algorithm ( GA ) is based on the stochastic technique, which can obtain the global optimal solution by effectively searching the whole solution space. But the Normal Genetic Algorithm (NGA) can't guarantee the speed of the search and the precision of the optimal solution simultaneously when the solution space is large. Double Population Genetic Algorithm (DPGA) is proposed to overcome this disadvantage of NGA. It takes advantage of the incorporation of two populations with different search strategy to guarantee the searching speed as well as the optimal solution precision. More than 50% computation time can be saved by applying DPGA to auto - matching interpretation of numerical well testing than using NGA.