5results about How to "Improve collaboration" patented technology

The application discloses a kind of off-site computing power dynamic collaborative optimization methods based on multi-agent reinforcement learning, including the following steps: constructing resource topology diagram;Based on resource topology diagram, get node-level state feature vector and system-level state feature vector;Multi-agent environment is constructed;Local observation, global information and agent action set are input into improved CTDE model, output policy network parameters and value network parameters and construct training batch;Based on training batch, get converged policy network parameters and converged value network parameters;Get execution result;Obtain dynamically updated policy network parameters and dynamically updated value network parameters, realize the dynamic collaborative optimization of task acceptance, resource allocation, task migration, replica start-stop and bandwidth matching.

This application discloses an external data repair method, apparatus, device, and storage medium for data lakes, relating to the fields of financial technology and data engineering technology. The method includes: acquiring scheduling information and reference relationship information of external data in the data lake; determining corresponding downstream protection link information based on the scheduling and reference relationship information; monitoring corresponding data entry start operations and abnormal failure operations based on the downstream protection link information to determine the data entry start operation information and abnormal failure operation information; identifying emergency handling tasks corresponding to the affected downstream operation links based on the data entry start operation information and abnormal failure operation information to determine emergency handling task information; and controlling the system based on the emergency handling task information to update or repair the data in the data lake. This application significantly improves response efficiency and ensures data lake security by accurately identifying affected downstream operation links and automatically generating emergency task control data repair.

The application relates to a 3D dispensing and mounting detection integrated system based on a dual-arm multi-vision cooperative robot, and belongs to the field of intelligent robots; a sensor tooling table is horizontally arranged on the upper surface of a workbench; a sensor detection table is arranged on one side of the sensor tooling table; a detection robot arm position corresponds to the sensor tooling table; a shock absorber tooling table and a chip placement position are adjacently arranged on the upper surface of the workbench; a dispensing robot arm is installed on one side of the shock absorber tooling table; a dispensing device is installed at the top end of the dispensing robot arm; a multi-vision visual assembly and a height sensor position correspond to the dispensing robot arm; a glue cleaning device is installed on one side of the shock absorber tooling table; the application realizes accurate positioning of parts and accurate detection of dispensing effect through the visual assembly, can effectively improve the production speed of products, and guarantees the qualified quality of products; moreover, the system guarantees stable taking and placing of the parts through the mechanical clamps driven by the robot and the air cylinder, realizes steady execution of the overall process, and improves the flexibility and cooperation of assembly.

This invention discloses a multi-robot on-orbit maintenance task allocation method based on Q-learning, comprising the following steps: Step 1, rasterizing the simulated on-orbit maintenance task space into a three-dimensional array model; Step 2, initializing the positions of the end effectors of each robot arm and the positions of multiple target points for maintenance tasks; Step 3, based on the Q-learning algorithm, obtaining the ideal paths for each robot arm that start simultaneously, achieve the optimal time to complete all tasks, and maximize the cooperation among the robots; Step 4, obtaining the task allocation and execution order of each robot arm based on its ideal path. This invention effectively solves the problem of unclear task allocation in multi-robot, multi-target scenarios, improves the cooperation between robots, reduces task completion time, and thus leverages the advantages of multiple robots.

This invention discloses a hierarchical dynamic scheduling method based on the policy optimization DDQN algorithm for solving the dynamic scheduling problem of a reentrant hybrid pipelined workshop with batch processing machines. First, the objective function and constraints of the scheduling problem are determined. Then, by introducing a self-attention mechanism, a hierarchical structure based on DDQN is proposed, constructing two agents: a batching agent and a scheduling agent, to solve the batching and scheduling subproblems respectively. Furthermore, to address the multi-stage batch processing and reentrant scheduling characteristics of the problem, a Markov decision process considering the characteristics of these two agents is designed, including state, action, and reward. Further, an action selection strategy based on masking combined with an ε-greedy strategy and a soft-start target network update strategy are proposed to improve efficiency and generalization ability. This invention demonstrates significant effectiveness in solving the dynamic scheduling problem of a reentrant hybrid pipelined workshop with batch processing machines.