Resource scheduling method and system based on multi-modal cooperation and deep reinforcement learning

By constructing a digital twin environment and a deep reinforcement learning model at the station, and integrating visual and auditory data to predict passenger flow trends, the optimal resource scheduling strategy is generated, solving the problems of slow response speed and poor decision consistency in the station resource scheduling system, and achieving efficient resource scheduling and energy consumption optimization.

CN121809984AActive Publication Date: 2026-04-07SUZHOU BOYUAN RONGTIAN INFORMATION TECH CO LTD
6 Cites 2 Cited by

Patent Information

Application Number
CN202610269173.6
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-03-06
Publication Date
2026-04-07
Estimated Expiration
2046-03-06

AI Technical Summary

Technical Problem

The existing station resource scheduling system is unable to adapt to the complex and changing passenger flow environment, resulting in problems such as slow resource scheduling response speed, low traffic efficiency, energy waste and poor decision-making consistency.

Method used

By constructing a digital twin simulation environment for the station, a multimodal state matrix is ​​generated by integrating visual passenger flow density and auditory environmental noise. A deep reinforcement learning model is used to predict passenger flow diffusion trends, generate a collaborative strategy for broadcast volume gain and directional screen path pointing, and perform physical space verification to finally generate the optimal resource scheduling instruction.

Benefits of technology

It enables rapid response to complex changes at stations, dynamically balances traffic flow efficiency and energy consumption, eliminates discrete instructions, and enhances decision-making consistency and on-site guidance synergy.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The invention relates to the technical field of resource scheduling, in particular to a resource scheduling method and system based on multi-modal collaboration and deep reinforcement learning, and the method comprises the steps: constructing a digital twin environment, fusing the collected visual passenger flow density and auditory environment noise in real time, and generating a multi-modal state matrix; predicting a passenger flow diffusion trend and fusing the passenger flow diffusion trend with the multi-modal state matrix to generate a composite feature vector; inputting the composite feature vector into a deep collaborative scheduling model, and generating a primary collaborative strategy containing broadcast volume gain and guide screen path pointing by using a strategy network with double constraints of evacuation efficiency maximization and energy consumption minimization; performing secondary constraint verification on the primary cooperation strategy to determine an optimal resource scheduling instruction; and issuing an optimal resource scheduling instruction to drive the field broadcast and the guide screen to execute actions. According to the method, quick response can be realized, the grooming efficiency and the operation energy consumption can be dynamically balanced, and instruction discretization caused by independent decision is effectively eliminated, so that the decision consistency of resource scheduling is improved.
Need to check novelty before this filing date? Find Prior Art