Transmission model of risk elements neural network in entropy system based on simulated annealing.

Autor: Li, Cunbin, Ding, Jia, Pan, Zhangyi
Předmět:
Zdroj: Journal of Industrial & Production Engineering; Jul2017, Vol. 34 Issue 5, p375-381, 7p
Abstrakt: Uncertainty decrease is an important attribute during the risk elements transmission process, and the occurrence of risk elements may not be clear until all factors tend to be steady. To study the interrelation among risk elements and their occurrence mechanism, this paper makes an analogy between risk elements transmission effect and molecular motion of thermodynamics and defines the entropy system based on entropy theory to integrate the macroscopic transmission with the microcosmic entropy change. A hypothesis about the state of risk elements during the entropy reduction process is proposed. And a risk elements neural network is built to simulate the transmission effect with simulated annealing algorithm controlling the entropy reduction. With a practical example, the hypothesis is proved and the coming risk elements are forecasted. This shows the model with a fast convergence rate effectively combines risk elements and entropy, which can be widely applied in the risk early-warning management. [ABSTRACT FROM AUTHOR]
Databáze: Complementary Index
Nepřihlášeným uživatelům se plný text nezobrazuje