【作者】 张海燕；

【导师】 王玉学；

【作者基本信息】 大庆石油学院 ， 计算机应用技术， 2008， 硕士

【摘要】 人工鱼群算法是将基于行为的人工智能思想通过动物自治体的模式引入到优化问题中,而提出的一种智能优化算法。该算法只需要比较目标函数值,对目标函数的性质要求不高,对初值的要求也不高,具有全局寻优的能力,寻优速度较快。油田注水系统运行优化问题是在整个系统的服务要求给定的情况下,确定注水泵的开启状态及其运行参数,使其满足系统配注要求,并能达到降低能损耗的目的。油田注水系统运行优化问题是一个含有0-1整型变量和连续变量的混合变量的优化问题,涉及的变量数目较多,一般可达到几十个甚至上百个,等式和不等式约束条件的数目多达几百个甚至上千个,因此对其进行求解比较困难。由于油田注水系统运行优化问题的混合变量使得问题的求解比较困难,本文研究了只以注水站的总流量为设计变量的优化问题。采用了改进的模式搜索算法以及改进的人工鱼群算法进行求解,并对两种算法进行了比较分析。本文对人工鱼群算法进行了改进,用于求解含有0-1整型变量和连续变量的混合变量优化问题。设计了一种两层人工鱼群算法,用内外两层人工鱼的状态分别表示0-1整型变量和连续变量,一种一层人工鱼群算法,只用一层人工鱼的状态表示0-1整型变量和连续变量,将两种不同的改进算法用于求解同一个注水系统运行优化问题,并对两种改进算法进行比较和分析。油田注水系统运行优化是含有约束条件的优化问题,由于约束条件的存在,使得优化算法往往因为约束条件的限制而寻找不到最优解或者没有找到最优解而直接退出。本文采用了罚函数方法将该问题转化为无约束优化问题,并结合改进的人工鱼群算法对转化后的无约束优化问题进行了求解。更多还原

【Abstract】 The artificial intelligent that based on animal’s behaviors can be used for the solving of optimization problems by animats’mode. The artificial fish algorithm is one of intelligent optimization algorithms which based on it. The algorithm just compares the values of object functions, so it has no special requirement for the problems. And the algorithm is insensitive to the initial values, has global search abilities for optimization and fast optimizing speed.The optimization problem of oilfield water system is to determine the open position and operation parameter of water injection stations under the instance that the service requires of the oilfield water system is given, in order to satisfy the requirement of the whole system and to cut down the power consumption. The optimization problem of oilfield water system is a non-linear optimization problem with mixed variant which has 0-1 integer variable and continuous variable. The problem has many variants, may be decades or hundreds, the number of equality constraint and non-equality constraint may be hundreds or more than thousands, so it is difficulty to solve the problem.The problem of the oilfield water system with mixed variant is difficult to solve. The article studies the optimization problem which is makes the flux of water injection stations as variant. We solve the problem with improved pattern search method and improved artificial fish algorithm, then compare and analyze the results.The article improves the artificial fish algorithm to solve the mixed variant optimization problem with 0-1 integer variable and continuous variable. The article design two algorithm, one of them is double-layer artificial fish algorithm that the fist layer artificial fish makes 0-1 integer variable the as variant and the second layer artificial fish makes continuous variable as variant. Another is one layer artificial algorithm which is makes 0-1 integer variable and continuous variable as variant. We solve the same problem of oilfield water system with two improved artificial fish algorithm, and then compare and analyze the two improved methods.The problem of oilfield water system has constraint conditions. The optimization algorithm will be exit without find the optimal solution because of the constraint conditions. The article uses the penalty function method transform the constraint optimization problem into non-constraint optimization problem, and then solves the non-constraint optimization problem with artificial fish algorithm.更多还原