论文部分内容阅读
油品调和是生产成品油的关键环节,调和配方决定了利润的最终大小。配方优化模型中物料平衡约束、产品产量约束属于线性约束,但质量指标约束中存在的非线性约束,使整个求解问题成为非线性约束求解问题。针对炼油厂油品调和复杂非线性约束配方优化问题,提出了一种新的线性先行GA算法。通过将线性约束与非线性约束拆分开,先求解凸多面体的全部顶点和极方向,解出满足线性约束的所有可行解所在的可行区域,进而用GA算法在此可行域内,通过优化凸组合的系数,从而达到搜索整个线性解区间的目的,完成非线性优化模型的求解。仿真结果表明,新算法大大缩小了GA算法的搜索空间和需要处理的约束,能快速的获得理想的调和配方,使利润最大并保证对质量卡边的要求。
Oil blending is the key to the production of refined oil, blending formula determines the final size of the profit. The material balance constraint in the formulation optimization model and the product yield constraint belong to the linear constraint, but the nonlinear constraint existing in the quality index constraint makes the entire solution problem become a nonlinear constraint solution. Aiming at the optimization problem of complex non-linear constrained recipes for oil blending in refinery, a new linear prior GA algorithm is proposed. By separating the linear constraint from the nonlinear constraint, all the vertices and polar orientations of the convex polyhedron are solved first, then the feasible regions where all the feasible solutions that satisfy the linear constraint are solved. Then, by using the GA algorithm in this feasible region, So as to achieve the purpose of searching the entire linear solution interval and to complete the solution of nonlinear optimization model. The simulation results show that the new algorithm greatly reduces the search space and the constraints that need to be dealt with, and can quickly obtain the ideal harmonic formula to maximize the profit and ensure the quality card edge.