Loading...
Please wait, while we are loading the content...
Similar Documents
Using neural networks for goal driven simulation
| Content Provider | Semantic Scholar |
|---|---|
| Author | Clavijo, Maria F. |
| Copyright Year | 2004 |
| Abstract | OF THE THESIS USING NEURAL NETWORKS FOR GOAL DRIVEN SIMULATION by Maria F. Clavijo Florida International University, 2004 Miami, Florida Professor Martha A. Centeno, Major Professor An integration framework for Neural Networks (NN) and Goal Driven Simulation (GDS) has been designed. It offers no constraints regarding number of variables (n>3) and it does not have domain restrictions. The effectiveness of the framework was tested by observing the computational time required for obtaining responses and for training, and by assessing its accuracy for different scenarios. This framework has achieved the automation objective set by GDS under a shorter time frame, as it reduces the time from more than 42 hours to less than 14. A trained NN generates responses to queries almost instantaneously. However, it requires time re-building and re-training new NNs when changes are made to the system represented by the model. If these changes are rare, the payoff is worthy as this approach gives users more flexibility. |
| File Format | PDF HTM / HTML |
| DOI | 10.25148/etd.FI14060853 |
| Alternate Webpage(s) | https://digitalcommons.fiu.edu/cgi/viewcontent.cgi?article=3569&context=etd&httpsredir=1&referer= |
| Alternate Webpage(s) | http://digitalcommons.fiu.edu/cgi/viewcontent.cgi?article=3569&context=etd |
| Alternate Webpage(s) | https://digitalcommons.fiu.edu/cgi/viewcontent.cgi?article=3569&context=etd |
| Alternate Webpage(s) | https://doi.org/10.25148/etd.FI14060853 |
| Language | English |
| Access Restriction | Open |
| Content Type | Text |
| Resource Type | Article |
