A discrete-based production process task handover management method and system

By constructing a process dependency graph and a hierarchical intelligent agent architecture, and combining delayed feedback signals and policy distillation mechanisms, the inherent logical relationship between process dependencies and handover decisions is solved, achieving highly efficient and adaptable automated management of production process task handover.

CN121882640BActive Publication Date: 2026-06-02杭州友成科技有限公司
View PDF 2 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
杭州友成科技有限公司
Filing Date
2026-03-19
Publication Date
2026-06-02

AI Technical Summary

Technical Problem

Existing production process task handover management methods are unable to effectively integrate the inherent logical connection between process dependencies and handover decisions, resulting in the inability to achieve highly efficient and adaptable automated management.

Method used

A process dependency graph is constructed, and abnormal pattern category labels and sampling weights are generated through clustering. Combined with the global scheduling agent, process group agent and operator agent in the hierarchical agent architecture, the delay reward value is calculated based on the delay feedback signal, and the meta-policy is passed to each layer agent through the policy distillation mechanism to dynamically update the policy parameters.

Benefits of technology

It improves the intelligent decision-making level and system adaptability of task handover in production processes, realizes the accuracy of delayed reward calculation and the rationality of decision evaluation, and optimizes the efficiency of policy knowledge transfer and learning adaptability in the hierarchical intelligent agent architecture.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN121882640B_ABST
    Figure CN121882640B_ABST
Patent Text Reader

Abstract

The present application relates to the technical field of production scheduling, and provides a production process task handover management method and system based on a discrete type, which comprises the following steps: collecting handover data to construct a process dependency graph containing path weights; clustering rejection reason codes to obtain an abnormal pattern and calculate a sampling weight; constructing a hierarchical intelligent agent containing global scheduling, process groups and operators based on production constraints and generating a meta-strategy; receiving a delay feedback, calculating a delay reward according to the path weight, backtracking the allocation, forming an experience sample, sampling and constructing a training batch according to the sampling weight; transferring the meta-strategy and updating the strategy parameters of each layer through strategy distillation; and generating a handover decision by the operator intelligent agent in combination with the current state. The present application improves the intelligent decision level and system adaptability of the production process task handover, and realizes continuous performance optimization through a complete delay reward allocation and strategy updating mechanism.
Need to check novelty before this filing date? Find Prior Art