基于鱼群算法的油田多级站定位优化方法研究

为了解决油田多级站定位优化问题,建立了该问题的大规模非线性MIP模型,该模型用传统方法求解相当困难.为了方便鱼群算法对该问题的求解和提高解算速度,对模型中的连续实型变量进行离散化处理,从而使整个优化模型变成纯0-1非线性IP模型.在解算过程中,用人工鱼体能累计和消耗程度来调度人工鱼行为;用海明距离度量人工鱼个体间的距离;采用随机步距移动的贪婪法描述个体追尾行为;采用鱼群规模、视野大小、拥挤程度和最低生存体能控制等方法实现局部最优解逃逸策略;采用最大迭代次数和迭代过程中最优解平均值变化程度来控制迭代终止时机.算例结果表明,该算法计算速度和稳定性有较大提高,可在微机上稳定地获取问题的最优解.

An optimization approach to the location of oilfield multistage stations by fish-swarm algorithm

A large-scale nonlinear MIP model for the locating optimization of oilfield multistage stations is established,but it is very difficult to solve it by traditional methods.To provide the convenience for solving the model using fish-swarm algorithm and enhance solving speed,the continuous real variables in the model are changed into discrete 0-1 variables so that the nonlinear MIP model is transferred into a pure 0-1 nonlinear IP model.In the solution process,the behaviors of a fish are determined by its body energy status;Hamming distance is used to measure the distance between two fishes;the following behavior of the fishes is described by the greedy method with random stepdistance;the number,the visual scope,the crowding degree and the lowest survival body energy of fishes are controlled to realize the escaping policy from local optimum solutions;the termination of iteration is controlled by maximum iterating times and the varying rate of the average value of the optimum solutions during iteration.The application shows that the computing speed and stability of the algorithm are higher,and the optimum solution of the model can be gained on microcomputers.