基于混合蛙跳算法的成品油管网优化设计  

Optimal Design of Multi-product Pipeline Network by Shuffled Frog Leaping Algorithm

作　　者：吴华丽[1] 汪玉春[1] 陈坤明[1] 唐煌 曹军[3] 

机构地区：[1]西南石油大学,四川成都610500 [2]中国石油西部管道有限公司,甘肃洒泉735009 [3]中国石油天然气集团公司新疆时代石油工程有限公司,新疆克拉玛依834000

出　　处：《石油工程建设》2008年第1期14-16,共3页Petroleum Engineering Construction

基　　金：四川省高校重点学科建设资助项目 (SZD0416)

摘　　要：成品油输送管网的参数优化模型,属于非线性离散变量优化组合问题,且目标函数具有多峰性,往往存在局部极小点,容易陷入局部寻优。为了克服传统算法及以往非线性规划的不足,文章尝试用一种新兴的启发式算法——混合蛙跳算法(SFLA)来解决此类参数优化难题。建立以管道建设费用为目标函数和以管道各节点流量方程、环路能量平衡方程等为约束条件的数学模型,采用SFLA算法进行求解。算例结果表明,SFLA算法可以较好地解决成品油输送管网的参数优化问题。同理,可应用于解离散域组合参数优化问题。The parameter optimal design model of multi-product pipeline network belongs to the optimal combination problem of non-linear discretization, and the objective function has the multiple peaks and some local minimum points, thus we easily run into the local optimal searching. For overcoming the deficiency of traditional algorithm and non-linear programming, this thesis has tried to solve the parameter optimal problem using a new heuristic algorithm—Shuffled Frog Leaping Algorithm(SFLA). Taking the flow rate equation at each node of the pipeline and the energy equilibrium equation as the constraint conditions and the construction expenses of pipeline as the objective function, the mathematic model is developed for the product oil pipeline network, then solved with SLFA .The result from calculational example shows that SLFA can solve the parameter optimal problem of the multi-product pipeline very well. Similarly, it can be applied to combined parameter optimal problems in discrete domain.

关 键 词：混合蛙跳算法 成品油管网 优化设计 数学模型 

分 类 号：TE973.1[石油与天然气工程—石油机械设备]