Performance evaluation method for multi-stage task system

Abstract

The embodiment of the invention provides a performance evaluation method for a multi-stage task system (PMS), which comprises the following steps of: on the basis of deep learning feature representation, fusing polymorphic feature data of the PMS on the basis of Graph feature input of a PMS damage evolution mechanism; according to task stage requirements of the PMS and a causal mechanism of Graph,applying factor graph segmentation to obtain an appropriate distributed evaluation framework; based on attention operator design of state & position contrast coding, acquiring a minimum Graph neighborhood module in distributed evaluation in a self-adaptive mode; and further focusing the target neighborhood, generating a GRU adaptive deep network capable of describing PMS stage correlation, and mining neighborhood information of the minimum evaluation sub-graph. By applying the scheme provided by the embodiment of the invention, the interpretability of the corresponding deep learning module isexpected to be increased, the complexity of the evaluation algorithm is greatly reduced, and the generalization ability of the model is improved, so that the performance of the PMS is accurately evaluated.

Description

technical field [0001] The invention relates to the field of computer technology, in particular to a multi-stage task system performance evaluation method and device. Background technique [0002] The multi-stage task system PMS refers to the complex system in which the whole system has multiple task stages, and each task stage performs different tasks. The working mode, environmental conditions, system configuration and failure criteria of the same component in the PMS will change with the change of the task phase. With the development of science and technology, modern systems are becoming more and more complex and intelligent. more important role. The performance of PMS not only affects the quality and economic utility of the complex system, but also affects the reliability of the equipment and the safety of the user, and may even cause huge human and material losses. [0003] However, due to the multi-stage task requirements in PMS, the model needs to consider cross-st...

AI Technical Summary

Problems solved by technology

The performance of PMS not only affects the quality and economic utility of complex systems, but also affects the reliability of equipment and the safety of users, and may even cause huge human and material losses.

[0003] However, due to the multi-stage task requirements in PMS, the model needs to consider cross-stage dependencies and intra-stage differences at the same time; moreover, the complexity of each component and the diversification of tasks in PMS require a large number of redundant designs

Therefore, it is difficult to quickly evaluate the performance of PMS

Images