Title :
Expert search in neural network optimization
Author :
Cheung, John Y. ; Chang, Chir-Ho
Author_Institution :
Sch. of Electr. Eng. & Comput. Sci., Oklahoma Univ., Norman, OK, USA
Abstract :
Recent research has shown that many optimization problems can be solved using artificial neural networks (ANN). The primary problem encountered in finding the globally optimal solution is the existence of local minima. The approach taken in this research is to develop a constrained optimization strategy using the observed data during adaptation or search. The constraints are rigidly enforced by an expert system so that the solution will always be a valid one
Keywords :
expert systems; neural nets; optimisation; artificial neural networks; expert system; globally optimal solution; local minima; neural network optimization; observed data; optimization problems; Artificial neural networks; Cities and towns; Constraint optimization; Equations; Expert systems; Intelligent networks; Neural networks; Neurons; Simulated annealing; Traveling salesman problems;
Conference_Titel :
Circuits and Systems, 1990., Proceedings of the 33rd Midwest Symposium on
Conference_Location :
Calgary, Alta.
Print_ISBN :
0-7803-0081-5
DOI :
10.1109/MWSCAS.1990.140644
