Methods, apparatuses, devices, and media for resource allocation

By constructing a multi-dimensional feature vector of users and using a resource quota determination model trained by reinforcement learning, the resource allocation of the sharing economy platform is dynamically optimized, solving the problem of poor user experience caused by fixed allocation strategies, realizing personalized and dynamically balanced resource allocation, and improving user satisfaction and platform efficiency.

CN122175193APending Publication Date: 2026-06-09BEIJING BAIJU YIXING TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN Β· China
Patent Type
Applications(China)
Current Assignee / Owner
BEIJING BAIJU YIXING TECH CO LTD
Filing Date
2026-01-27
Publication Date
2026-06-09

AI Technical Summary

Technical Problem

In the sharing economy platform, the fixed number of resources allocated cannot meet the needs of workers with high resource acquisition, and newly joined workers or workers with low resource acquisition cannot make full use of the allocated resources, resulting in a poor user experience.

Method used

By acquiring user data, constructing multi-dimensional feature vectors for users, and using resource quota determination models trained based on reinforcement learning, the amount of resources allocated is dynamically optimized. Combining users' historical resource allocation characteristics, remaining resource characteristics, user level, and operating environment characteristics, personalized resource allocation strategies are generated.

Benefits of technology

It enables personalized and dynamic balancing of user resource allocation, improves user satisfaction, enhances platform operational efficiency, and avoids resource mismatch issues caused by static rules.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122175193A_ABST
    Figure CN122175193A_ABST
Patent Text Reader

Abstract

This application discloses a method, apparatus, device, and medium for resource allocation, relating to the field of resource allocation technology. The method includes: responding to a resource allocation instruction issued by a user, acquiring user data, determining multi-dimensional user characteristics based on the user data, wherein the multi-dimensional user characteristics include the user's historical resource allocation characteristics and the user's remaining resource characteristics; concatenating the features from the multi-dimensional user characteristics to obtain a user state vector, inputting the user state vector into a resource quota determination model to generate the user's resource allocation quantity, wherein, based on the resource quota determination model, calculating the expected reward corresponding to the resource allocation quantity of the user state vector, and determining the user's resource allocation quantity based on the resource allocation quantity corresponding to the maximum expected reward. This avoids the problems of static rules or fixed-ratio resource allocation quantities, which are detached from the user's historical resource allocation quantities and remaining resource quantities, and can improve user satisfaction with resource allocation.
Need to check novelty before this filing date? Find Prior Art