【摘要】 根据现有炼油厂原油库存调度优化大多基于线性模型,通过混合整数线性规划来求解的情况,指出这种模型仅将复杂的生产操作逻辑关系进行简单的线性化,而没有考虑各种逻辑关系的内在联系,存在模型表达复杂、优化算法效率不高等问题。对此介绍了采用专家系统对上述的逻辑关系进行特定推理后再进行优化计算,经过反复迭代而构成的一种混合算法。应用该算法对实例的计算结果表明,它不仅具有模型表达简单的优点,而且效率较单纯的混合整数规划算法也有了一定程度的提高。更多还原

【Abstract】 At present, most of crude inventory scheduling optimization problems in refinery are solved by mixed integer linear programming (MILP) based on linear models. Because such models only simply have operation rules linearized and do not take account the internal connection between them, these methods usually have such drawbacks as complex procedure for model building, lower computation efficiency, etc. In this paper, one hybrid algorithm is proposed, during each iterative step of which the logic relationships mentioned above are deduced through expert system before optimization. The computation result showed that such an algorithm, compared with MILP, not only took fewer computation times, but also gave model in a more concise manner.更多还原