Complex product autonomous construction method and module based on man-machine hybrid enhancement

Abstract

The embodiment of the invention discloses a complex product autonomous construction method and module based on man-machine hybrid enhancement, and the method comprises the steps: S10, building a parameterized knowledge model from a key feature state to action mapping according to the knowledge of a person, and enabling a to-be-tested parameter in the knowledge model to be coded into an output vector of a neural network; S20, encoding an input vector of the neural network according to the key feature state of the environment and the reward function output value; S40, configuring an intelligentagent; S50, generating training data from a key feature state to action mapping on the basis of known experience knowledge of people, reversely fitting neural network parameters by utilizing the training data, and driving a knowledge model to deduce in a training environment; and S60, continuously optimizing the neural network parameters based on reinforcement learning until the training of the neural network is completed, and finally realizing the self-adaption of the intelligent agent.

Description

Technical field [0001] The present invention relates to the field of complex products. More specifically, to a man-machine-based product enhanced mixing complex to build their own methods, modules, storage media and computer equipment. Background technique [0002] Complex product is a product of complex structure, design and manufacturing process complex, high-tech, knowledge and integration of cross involve a number of disciplines, capital-intensive, technology-intensive and knowledge-intensive products, complex product design is in many fields of mechanical, electronic, control, etc. concurrent design process. [0003] Traditionally complex aerospace products, represented mainly by the complexity of product design human experience and knowledge, combined with computational analysis and simulation and other means, top-down level by level index system of decision-making, technical pathway analysis, the overall program design. With the complex technical requirements of a new gene...

AI Technical Summary

Problems solved by technology

With the new generation of complex products requiring higher and higher technical indicators, more and more requirements for working conditions, and more and more difficult requirements for game cooperation, it has become difficult to close, optimize, and speed up the design. Traditional methods have already It is increasingly difficult to adapt to the development of a new generation of complex products

[0004] In essence, the complex product design problem is a multi-objective and multi-constraint optimization problem for high-order multi-dimensional nonlinear systems with high-difficulty technical indicators, complex working conditions, and multi-party game cooperation characteristics

Limited by human's lack of abstraction level, association ability and learning efficiency for complex problems, there are "difficulty in abstraction" for models of high-order multi-dimensional nonlinear systems, "difficulty in association" for feasible solutions under multi-objective and multi-constraint requirements, and "difficulty in association" for Quantitative laws of existing analysis and test data are "difficult to learn" and other problems

Images