Другие журналы
|
scientific edition of Bauman MSTUSCIENCE & EDUCATIONBauman Moscow State Technical University. El № FS 77 - 48211. ISSN 1994-0408
The Comparison of Methods for Genetic and Variational Genetic Programming Using a Control Synthesis Problem for the Model "Predator-Victim" as an Example
# 05, May 2014 DOI: 10.7463/0514.0709252
Article file:
Diveev_A.pdf
(901.87Kb)
This paper studies numerical methods to solve the problem of control system synthesis. Two symbolical regression methods are compared such as a genetic programming method and a new variational genetic programming method. First, the paper provides the qualitative comparative analysis of methods and gives the numerical estimations of resources to be used by the methods. Then the methods are applied to solve a control system synthesis problem. The non-linear model of the "predator-prey" system is considered. To obtain the reliable results an identical values were assigned to the initial parameters of the search algorithms for both methods. A trivial basic solution is chosen for the variational genetic programming method, though the main advantage of this method is the possibility to reduce time of calculations owing to experienced developer's choice of the basic solution that is close to an optimal one. We performed a computational experiment to show that the variational genetic programming method finds a many times faster solution and spends less resources, and a received result provides better values of the goal functions, than the genetic programming method. In the future it is necessary to compare these methods using a complex challenging control system synthesis problem where it can be proven that the method of variational genetic programming has an advantage arising from the choice of the specific basic solution. It is also necessary to compare the method of variational genetic programming with other symbolic regression methods, such as analytical programming, grammatical evolution and network operator methods. References
Publications with keywords: synthesis of control system, the method of variations of the basis solutions, variational genetic programming Publications with words: synthesis of control system, the method of variations of the basis solutions, variational genetic programming See also: Thematic rubrics: Поделиться:
|
|
|||||||||||||||||||||||||||||
|