This paper relates to the field of refined oil
resource scheduling technology, and provides a method, apparatus, equipment, and storage medium for refined oil
resource scheduling. The method includes: using a simulator to simulate the scheduling scheme for the first sub-cycle of the
current cycle based on the node parameters of each node, to obtain the
inventory level and sub-cycle
reward value of each node at the end of the first sub-cycle; using a deep
reinforcement learning module to predict the target
inventory level of each node at the end of the first sub-cycle of the next cycle based on the
inventory level, sub-cycle
reward value, and scheme penalty value of the
current cycle, with the second sub-cycle of the
current cycle as the first sub-cycle of the next cycle; using a mathematical
programming module to determine the scheduling scheme for the next cycle based on the target inventory level and the operational constraints of each node, with the goal of minimizing scheduling costs and inventory level differences; and performing refined oil
resource scheduling on each node according to the scheduling scheme of the first sub-cycle of the next cycle. The embodiments described in this paper can improve the efficiency of refined oil resource scheduling.