A hybrid system integrating a piecewise linear representation and a neural network for stock prediction

2011 ◽  
Author(s):  
Yung-Keun Kwon ◽  
Hui-Di Sun
Keyword(s):  
Neural Network ◽  
Hybrid System ◽  
Piecewise Linear ◽  
Linear Representation ◽  
Stock Prediction ◽  
Piecewise Linear Representation

An Alternative Design Approach for Vehicle Crash Safety Using Crash Pulse Optimization

2000 ◽  
Author(s):  
R. J. Yang ◽  
C. H. Tho ◽  
C. C. Gearhart ◽  
Y. Fu
Keyword(s):  
Time Domain ◽  
Numerical Optimization ◽  
Piecewise Linear ◽  
Linear Representation ◽  
Optimization Techniques ◽  
Safety Requirements ◽  
Crash Safety ◽  
Design Variables ◽  
Piecewise Linear Representation ◽  
Occupant Injury

Abstract This paper presents an approach, based on numerical optimization techniques, to identify an ideal (5 star) crash pulse and generate a band of acceptable crash pulses surrounding that ideal pulse. This band can be used by engineers to quickly determine whether a design will satisfy government and corporate safety requirements, and whether the design will satisfy the requirements for a 5 star crash rating. A piecewise linear representation of the crash pulse with two plateaus is employed for its conceptual simplicity and because such a pulse has been shown to be sufficient for reproducing occupant injury behavior when used as input into MADYMO models. The piecewise linear crash pulse is parameterized with 7 design variables (5 for time domain and 2 for acceleration domain) in the optimization process. A series of sample runs are conducted to validate that pulses falling within the acceptable crash pulse band do in fact satisfy 5 star requirements.

Sign in / Sign up

Export Citation Format

Share Document