The application discloses an
inventory management method and
system based on real-time demand prediction, and the
system comprises a plurality of inventory nodes, a management module and a
server. Each inventory node acquires an
inventory data periodic change rate through a first
data acquisition module and acquires a demand data periodic change rate scheduled to each coverage area through a second
data acquisition module. The
server generates a predicted demand sequence and a sequence
uncertainty interval thereof based on the change rates, calculates a demand exchange rate between nodes and an inventory
demand rate of each node, determines the demand exchange rate according to the
negative correlation of the predicted demand sequence, and determines the inventory
demand rate according to the predicted demand and the
uncertainty interval. The management module performs commodity exchange between nodes according to the demand exchange rate until the inventory
demand rate reaches a preset value. The application realizes the collaborative scheduling of multi-node and multi-area inventories through the change rate driven demand prediction,
uncertainty quantification and node complementary matching, and reduces the
inventory cost.