EvoSort - System for Evolutionary Development of Generic Sorting Networks

EvoSort - System for Evolutionary Development of Generic Sorting Networks

Software

The system is intended for automatic design of algorithms for the construction of arbitrarily large sorting networks by means of genetic algorithm and continual instruction-based development. The input of the application consists of development parameters of the sorting networks, an embryo (initial instance of sorting network) parameters of genetic algorithm. After the correct settings and successful termination of the evolutionary design process, a sequence of instructions (a program) represents a result of the application whose iterative application on the embryo allows to design sorting networks satisfying the specified parameters. The resulting programs need to be verified in order to determine the ability of development of generic sorting networks. The application is able to analyze and visualize the resulting sorting networks. A detailed description of the approach is given in http://www.fit.vutbr.cz/research/view_pub.php?id=9459.

By downloading EvoSort (see also the link above) the user expresses agreement with the license terms.

The system is intended for automatic design of algorithms for the construction of arbitrarily large sorting networks by means of genetic algorithm and continual instruction-based development. The input of the application consists of development parameters of the sorting networks, an embryo (initial instance of sorting network) parameters of genetic algorithm. After the correct settings and successful termination of the evolutionary design process, a sequence of instructions (a program) represents a result of the application whose iterative application on the embryo allows to design sorting networks satisfying the specified parameters. The resulting programs need to be verified in order to determine the ability of development of generic sorting networks. The application is able to analyze and visualize the resulting sorting networks. A detailed description of the approach is given in http://www.fit.vutbr.cz/research/view_pub.php?id=9459.

By downloading EvoSort (see also the link above) the user expresses agreement with the license terms.

Genetic algorithm, instruction, development, sorting network.

Genetic algorithm, instruction, development, sorting network.

http://www.fit.vutbr.cz/research/view_product.php?id=178¬itle=1

Use of the result by another entity is possible without acquiring a license in some cases

http://www.fit.vutbr.cz/research/view_product.php?id=178¬itle=1

EvoSort.tar