Reinforced learning method for dynamically optimizing logistics scheduling and path planning in machining process

Abstract

The invention discloses a reinforcement learning method for dynamically optimizing logistics scheduling and path planning in a machining process, and belongs to the field of intelligent manufacturing.The method comprises the following steps: constructing a total state matrix of logistics scheduling at a certain moment as the input of the neural network; constructing a reward and punishment valueaccording to the total time length of all operation units waiting for materials in the primary processing process and the total number of conflict times of the AGVs, establishing a loss function of aneural network according to the reward and punishment value, training the neural network to maximize the reward and punishment value obtained in the final primary processing process, and obtaining a reinforcement learning decision model; and then for a new processing task, establishing a total state matrix at the current moment in real time, and inputting the total state matrix into the reinforcement learning decision model to obtain an action needing to be executed by the AGV at the current moment. According to the method, autonomous optimal path planning of the AGV in the machining process can be realized, and the technical problem that the logistics scheduling scheme cannot be dynamically and automatically adjusted in real time in the machining process in the prior art is solved.

Description

technical field [0001] The invention belongs to the field of intelligent manufacturing, and more specifically relates to a reinforcement learning method for logistics scheduling and path planning in a dynamically optimized processing process. Background technique [0002] In processing-intensive and discrete manufacturing fields such as molds and lithium battery equipment, the manufacturing process includes a series of complex, multi-process, and coupled processing technologies, which are customized, small-batch, large-scale, and multi-variety manufacturing. During its processing, resources such as materials, components, and equipment are diversified and have a wide range of dynamic characteristics. Therefore, the management and supply efficiency of material resources in the processing process directly affect the efficiency of the manufacturing process, which cannot be ignored. [0003] In the logistics scheduling process, the AGV trolley performs the transportation task of...

AI Technical Summary

Problems solved by technology

[0005] In view of the above defects or improvement needs of the prior art, the present invention provides a reinforcement learning method for dynamic optimization of logistics scheduling and path planning in the process of processing. structure and dynamic characteristics, build and train the reinforcement learning decision-making model, and realize the autonomous optimal path planning of the AGV trolley in the processing process, thereby solving the problem that the scheduling model in the prior art ignores the dynamic characteristics of the multi-process processing process and cannot be used in the processing Technical issues of real-time dynamic automatic adjustment of logistics scheduling scheme in the process

Images