next up previous
Next: How to program using Up: GAGSa flexible Previous: A new framework for



GAGS  library, or the application generator, has been applied in its previous versions to a variety of problems.

Optimization of a fuzzy classifier
In this application, [Rojas, 1996][Rojas et al., 1996], the fuzzy rules used by a controller are optimized. The linguistic values of the membership functions and the fuzzy rules conclusions are coded, and optimal values are found in an efficient way. GAGS allowed us to develop this application in a minimal amount of time; basically, only the fitness function had to be coded.

Optimal test pattern generation
Pattern generation can be modelled as an optimization problem in which the target is to determine a test signal that maximizes the quadratic difference between the nominal response of a circuit and the faulty one due to a defect. This application, presented in [Bernier et al., 1995] was also programmed in very few hours, and results obtained improved those obtained by the other algorithms tested: simulated annealing, a heuristic algorithm and a Hopfield neural network. Obviously, since many solutions are tackled at the same time, it takes much more time than any of them. This program has been included in the GAGS release as an example for the application generator.

Search for a solution in the Mastermind game
Playing Mastermind is a search problem, in which the player must find a hidden combination set by the oponent. The player lays a combination of tokens, and the oponent answers with the number of tokens in a correct position, without poiting out which ones they are, and the number of tokens with correct color. The player must find the correct combination based on these hints. In this case, presented in [Bernier et al., 1996], the GA looks for a combination that meets the restrictions made so far by the oponent, and plays it. It manages to play better than other algorithms, like simulated annealing and a random search algorithm, without the need to examine all the search space; besides, its performance does not degrade significantly with the size of the problem; the number of combinations needed to find the correct one increases roughly linearly with the size of the search space. This application uses fully the capabilities of the library; since it is a complex problem, the application generator could not be used. Besides, some special-purpose operators had to be applied.

Job shop scheduling
Production scheduling is a problem of optimizing the time and resources needed to accomplish a set of tasks while, at the same time, satisfying some constraints. In this work, published in this same proceedings [Gutiérrez et al., 1996], GAs are used to optimize the makespan of a set of operations. GAGS had to be adapted extensively to use repair operators.

next up previous
Next: How to program using Up: GAGSa flexible Previous: A new framework for

[JJ J. J. Merelo *[E-MAIL]
Equipo GeNeura -- GeNeura Team
Departamento de Electrónica y Tecnología de los Computadores
Universidad de Granada Granada (Espaņa)
Phone: +34-58-243162; Fax: +34-58-243230