Multi-scene performance index prediction method and system for semiconductor production line

Abstract

The invention relates to a multi-scene performance index prediction method and system for a semiconductor production line, and the system comprises a production scene quantitative dividing module which is driven by data, carries out the quantitative mapping of a product value, a product average processing period and the utilization rate of all equipment of the production line, and divides the production line into a light-load scene, a normal-load scene and a heavy-load scene; a main prediction network construction module which is used for constructing a prediction network in a normal load scene by taking the divided normal load data as sample data and combining a deep neural network algorithm with semiconductor production line performance prediction; and a multi-scene prediction model construction module which is used for applying the idea of transfer learning to production line prediction and constructing networks under light-load and heavy-load scenes according to the constructed main prediction network so as to form a multi-scene prediction model. Compared with the prior art, the production line scenes are quantitatively divided, multiple production line performance indexes under different load scenes can be more accurately predicted, and online use can be achieved.

Description

technical field [0001] The present invention relates to the technical field of intelligent manufacturing, in particular to a multi-scenario performance index prediction method and system for semiconductor production lines. Background technique [0002] The manufacturing industry is the main body of the national economy, the engine of rapid economic growth, and an important manifestation of comprehensive national strength. In order to ensure competitiveness, manufacturing enterprises need to quickly carry out dynamic scheduling management and allocate resources reasonably. Real-time performance index prediction results can provide a basis for production decision-making and evaluation, so it is an inevitable requirement to quickly understand the performance index achieved by the production line under a certain scheduling mode. [0003] At present, there are mainly three methods of predictive model modeling: mathematical modeling, simulation modeling, and machine learning mode...

AI Technical Summary

Problems solved by technology

At present, there are no literatures and patents related to the above-mentioned production line performance prediction method

Images