基于混合遗传算法的树状注水管网拓扑优化

以投资最小为目标函数,以井间隶属关系的惟一性、连接注水井数量等限制为约束条件,建立了树状注水管网拓扑优化数学模型.根据模型特点,将优化问题分为分配层和几何层,分别采用遗传算法和非线性优化方法进行求解.并对遗传算法的操作过程进行了改进,调整了适应函数,采用随机多父辈适应函数值加权交叉和多种变异操作,结合了模拟退火算法,并给出了初温的确定方法,改进了选择复制操作,从而形成了混合遗传算法.该算法能够有效地提高收敛速度,避免早熟收敛.利用算例说明了该方法的有效性.

opology optimization of tree-type water-injection pipe network based on hybrid genetic algorithm

A mathematical model for topology optimization of the tree type water injection pipe network was established by taking the minimum investment as the objective function and taking the restrictions to the unique relations between wells and stations, and numbers of connected water-injection wells as the constraint conditions. The optimization problem was divided into two layers of distribulion and geometry. The genetic algorithm and nonlinear optimization were used to solve the problems respectively. The operational process of genetic algorithm was improved, and the fitness function was adjusted. Combined with the simulation annealing algorithm,it adopted the random parent-number fitness-weighted cross and many mutation methods and gave the method of determining initial temperature. The selection operation was also improved. Then a hybrid genetic algorithm was given. This algorithm can improve the speed of convergence and avoid premature convergence. Example shows the efficiency of the algorithm.