Cloud-Based Resource Allocation Optimization in Multi-Enterprise Commerce and E-Commerce Platforms Using Deep Reinforcement Learning

Abstract

The paper outlines a new method of efficient resource distribution in multi-enterprise commerce and e-commerce systems, which has been achieved with the help of the Deep Reinforcement Learning. The suggested technique incorporates Min-Max Scaling and Z-Score Normalization to perform effective data preprocess to achieve faster convergence and better model stability. Moreover, Recursive Feature Elimination is used in order to eliminate the irrelevant features and improve the model performance. Deep Q-Network, an algorithm based on the use of a TensorFlow, is used to make decisions on how to allocate cloud resources in a dynamic manner, which is far more efficient than the conventional approach. The findings indicate that the proposed system optimizes the use of resource, minimizes the computational load, and real-time adaptation to changing demands and provide a scalable solution to large-scale e-commerce platforms. This study will advance AI-based solutions to cloud resource management and show that DRL can be used to improve the level of operational efficiency in the work of multi-enterprise complexes.