I have studied many research article related to guaranteed cost control problem with neural network. In these papers the authors find only the upper bound value of the cost function(guaranteed cost value). Recently the authors Z-P. Wang and H-N. Wu are studied the paper entitled on " Finite dimensional guaranteed cost sampled-data fuzzy control for a class of nonlinear distributed parameter systems". In this paper the authors calculated the value of the cost function and also find the upper bound of the cost function. I have a trouble with calculate the cost function value. How to find it?
please see the link:-
http://www.sciencedirect.com/science/article/pii/S0020025515005836
If you work with stochastic controlled infinite non-stationary Markov process you can use as cost function the fuzzy neural entropy. In Bayesian case it is not always Boltzmann-Gibbs entropy.
Well, the best way is to ask the authors of the mentioned articles.The cost function can be defrined in many ways. Some of them can have solution easily, such as the quadtaric function in genetic algorithm. However, to improve the chance of publshing the articles, some authors adopted a very mathematical cost function which can not have solution easily. Try your luck on this subject, or leave it.
- Badges
- Science topic