The project is to create client-server applications for BOINC. On the client side there should be Artificial Neural Network (ANN), trained by Genetic Algorithms (GA). The idea is ANN to be trained to play game Go. There is no need client side application to be developed. There are open source libraries which can be used as shown:
[login to view URL]
[login to view URL]
In fact [login to view URL] implements everything that client application should do (except distributed computing parameters which will come from BOINC server side).
On the server side there are group of applications:
1. Job submitter.
2. Task generator.
[login to view URL]
3. Result validator.
[login to view URL]
4. Result acceptor.
[login to view URL]
Each task consist of - Go board parameters, ANN topology information, sets of ANN weights (GA chromosomes) and parameters of GA optimizer (population size, generations number and etc.).
Each task result consists of - optimized set of ANN weights for particular Go board and ANN topology.
Most of the code is done but it needs compilation and validation.