Loading...
Please wait, while we are loading the content...
Similar Documents
Otimização dinâmica em tempo real utilizando modelos não-lineares simplificados
| Content Provider | Semantic Scholar |
|---|---|
| Author | Duraiski, Ricardo |
| Copyright Year | 2009 |
| Abstract | A new algorithm based on the goal attainment method is proposed. The algorithm, named GA-GRG, is a more appropriate way to settle multi-objective optimization problems. The algorithm is also based on the GRG method, with important changes to improve the convergence to the solution. The first modification consists of performing a factorization of the gradient of the constraints, allowing finding the most sensitive independent variables to the search direction, avoiding choosing inappropriate variables. Other important modification is the way in which the search direction is determined. In the proposed algorithm, the search direction is determined through an optimization problem with analytic solution applicable to any problem, independent on the nonlinearities of the constraint and objective functions. The third important modification of this algorithm consists of changing the original nonlinear problem composed by a set of nonlinear objective functions and nonlinear constraints into a problem with one linear objective function subject to nonlinear constraints, which allows the algorithm to rides only on the constraints (Active Set Method) of the modified system. The results show the high performance and high accuracy of the algorithm when applied to several optimization problems. |
| File Format | PDF HTM / HTML |
| Alternate Webpage(s) | https://lume.ufrgs.br/bitstream/handle/10183/16483/000702797.pdf;sequence=1 |
| Language | English |
| Access Restriction | Open |
| Content Type | Text |
| Resource Type | Article |
