Students' Software Project
"Parallel Implementation of Genetic Alhorithms"
(Antares)
"The Design And Implementation of Open System With Genetic
Algorithms And Multilayer Neural Network (Back Propagation) To Aproximate
Object Function"
Genetic algorithms represent one of the biologically inspired
paradigms used for solution of complicated optimisation problems. Genetic
algorithm works with Population (=set of acceptable solutions) New
generation is set in from old one by using operation of Selection, Crossover
and Mutation. The goal is to get the population with solution as good as possible.
The qality to a result gives the evaluation function. Genetic algorithms
dont need any special analytical property of this function. In case that the
function cant be expressed analitically, neural network will be used
for its aproximation.
Means: The system will be realised on workstations of UNIX operating
system. For parallel computing will be used PVM (Parallel Virtual Machine)
system. GUI (Graphical User Interface) is expected to be X-Windows.