Abstract: Considering the parameter fluctuations, using multi-objective weighting method, the present value sum of pipe and tank investment and maintenance, energy costs and mixed oil processing costs in life-cycle as the objective function, network layout situation as the decision variables, a product pipeline network optimal layout model was established. Use the genetic algorithm to solve it. This method is clear, fast and effective which leads to a satisfactory solution. Calculation results show that the optimal solution calculated by GA is more economic than the Kruskal or Prim algorithm, and product pipeline network parameter must be strictly set. If the parameters have a larger range, change the weight coefficients can generates a variety of designs for choosing. MATLAB genetic algorithm toolbox provides a friendly interface and easy to set which is convenient for engineering applications.

Key words: Genetic algorithm, Product oil, Pipeline network, Optimum layout, MATLAB

摘要： 考虑到成品油管网各参数的波动，采用多目标加权法，以生命周期内管道与罐区的投资与维护费用现值、能耗费用现值及混油处理费用现值之和为目标函数，管网布置情况为决策变量，建立了成品油树状管网优化布置的数学模型，并采用遗传算法进行求解。该算法原理清晰，能快速搜索到令人满意的有效解。计算结果表明:利用遗传算法所得的最优解比Kruskal或Prim算法更经济。成品油管网各参数的设定要严格准确，若参数波动范围较大，则可通过变换权重系数的方法生成多种设计方案以供选择。MATLAB提供的遗传算法工具箱界面友好、设置简便，适于工程应用。

关键词: 遗传算法, 成品油, 管网, 优化布置, MATLAB