Cross-domain collaborative workshop dynamic scheduling method based on balance index adaptive evolution

Abstract

The invention discloses a cross-domain collaborative workshop dynamic scheduling method based on balance index adaptive evolution, and relates to the technical field of workshop dynamic scheduling. The method comprises the following specific steps: performing discrete indirect coding on a to-be-scheduled task; after the discrete indirect coding is completed, traversing the coding so as to convert a sequence, and generating a new coding individual by adopting an operator based on an exchange sequence; and carrying out evaluation constraint on the coded individuals by adopting a self-adaptive evolution strategy based on a balance index, and taking an evaluation result as an index for parent individual selection and local search selection. According to the method, by designing a discrete indirect variable coding form, generation of an infeasible solution is avoided, so that the solving speed of the algorithm is improved; and an operator is generated based on the new solution of the exchange sequence, so that individuals can be quickly and efficiently generated. Meanwhile, in combination with a self-adaptive evolutionary strategy based on a balance index, the algorithm is stable in effect when being applied to scheduling problems in different scenes, and has a good solving speed.

Description

technical field [0001] The invention relates to the technical field of workshop dynamic scheduling, and more particularly relates to a cross-domain collaborative workshop dynamic scheduling method based on self-adaptive evolution of a balance index. Background technique [0002] Since different manufacturing fields have certain similarities, the complementary synergies of various fields in the distributed workshop environment make the cross-domain synergy effect good. However, tasks and environments may change dynamically and the complex characteristics of cross-domain collaboration scenarios make it difficult for traditional methods to achieve cross-domain collaborative distributed workshop dynamic scheduling. [0003] In the prior art, some methods consider the idea of ​​using traditional algorithm improvements combined with other algorithms for dynamic scheduling. Because this kind of algorithm simplifies the scene without considering the heterogeneity of related problem...

AI Technical Summary

Problems solved by technology

However, the existing reinforcement learning models are only applicable to specific scenarios; when tasks, resources, environments, and scheduling objectives change, the models are not applicable and need to be retrained for new scenarios, so dynamic scheduling algorithms based on reinforcement learning exist It is difficult to adapt to the problem, and reinforcement learning does not require prior knowledge, and requires a large amount of data support for model training, which makes model learning inefficient and takes a long time

Images