The article considers the problem of decreasing efficiency in classical hard-coded sorting systems under the conditions of exponential growth in e-commerce volumes. To solve this problem, the implementation of autonomous dynamic parcel distribution algorithms based on machine learning (ML) and artificial intelligence (AI) is proposed. The process of collecting primary cargo metrics using automated DWS (Dimensioning, Weighing, Scanning) systems and the routing algorithm that takes into account the dynamic load parameters of the logistics center are described. The architecture of the control algorithm's interaction with the physical mechatronic objects of the warehouse — stacker cranes (AS/RS), autonomous mobile robots (AMR), and sorting conveyors — through the WCS management system is analyzed. The requirements for the IT infrastructure, including Edge Computing technologies and the creation of Digital Twins, are substantiated. It is proven that the synergy of intelligent software and physical equipment transforms sorting centers into flexible adaptive ecosystems with maximum throughput capacity.