Abstract: Product oil blending is an important step for oil refining, which directly affects the economic benefits of oil refineries. Using petrochemical industry as background, this paper does a research on the optimization of the recipe for product oil blending. A mixed-integer programming model is developed to minimize the total production cost of the refinery on the condition that product quality indexes are satisfied, and an efficient solving strategy based on genetic algorithm is proposed. Then a simulation experiment is made according to the actual production data of an oil refinery and the balance relationship between the inventory cost and the setup cost is showed bycomputational results. To be specific, the total inventory cost increases and the number of setups decreases, when the unit inventory cost remaines constant and the unit setup cost increases. Conversely, the number of setups increases and the total inventory cost decreases, when the unit setup cost remaines constant and the unit inventory cost increases.

Key words: product oil blending, production-inventory planning, genetic algorithm, refinery industry

摘要： 成品油调和是石油炼制过程中的重要环节,直接影响炼油企业的经济效益。本文以石化行业为背景,针对成品油调和配方优化问题进行了研究,在满足成品油质量指标约束的条件下,以最小化企业生产成本为目标,建立了混合整数规划模型,提出了基于遗传算法的有效求解策略,并根据某炼油厂的实际生产数据进行了仿真实验,计算结果反映了库存成本与启动成本之间的平衡关系,即:当单位库存成本不变,单位启动成本逐渐变大时,库存总成本随之增大,启动次数随之减少。反之,当单位启动成本不变,单位库存成本逐渐变大时,启动次数随之增大,库存总成本随之减少。

关键词: 成品油调和, 生产库存计划, 遗传算法, 炼油工业