The D-MASON enviroment enables execution of distributed multi-agent simulation models. In cooperation with ISISLab Salerno University (Italy) I work on tool that will automate creation of distributed computing clusters in the Amazon Web Services cloud. The tool will be created as a plugin to the MIT StarCluster enviroment - one of the most popular HPC tools for the cloud.
SOF is a toolset that utilizes the map-reduce model along with the Hadoop platform for HPC computations in the cloud.
The source code along with the technical documentation can be obtained at https://github.com/isislab-unisa/sof.
A detailed project concept is described in the paper: M. Carillo, G. Cordasco V., Scarano, F. Serrapica, C. Spagnuolo., P. Szufel, SOF: Zero Configuration Simulation Optimization Framework on the Cloud, In Proc. of 24th Euromicro International Conference on Parallel, Distributed and Network-Based Processing, Heraklion,Crete, 2016.
I have created a bash temaplate that enables rapid computing cluster creation in the cloud.
Starting is simple - just create any Ubuntu instance (e.g. t2.micro) in the Virginia region, copy
credentials.csv file to
/home/ubuntu and run the following commands
wget https://szufel.pl/sc_setup.sh bash sc_setup.sh
"Raising Open and User-friendly Transparency-Enabling Technologies for Public Administrations" (ROUTE-TO-PA) is a reasearch project financed by the 8th EU Framework Programme - Horizon 2020
ROUTE-TO-PA is an innovation project aimed at creating tools and software for increasing transparency of public administration.
The goal of the Decision Support and Analysis Division team is to create a multi-agent simulation tool to analyse preference elicitation in heterogenous communities.
The project web site is available at http://www.routetopa.eu/
In cooperation with M. Myślak, B. Kamiński and M. Jakubczyk we have created SilverDecisions applications.
The software allows for creation and analysis of decision trees.
The application can be used in web browser at http://silverdecisions.pl.
The develpmnent page of the project is avaliable on BitBucket.
Asynchronous Knowledge Gradient algorithm enables distributed stochasic optimization computations for ranking & selection problems.
Implementation in Java language of the Asynchronous Knowledge Gradient algorithm can be downloaded here.
Documentation of the algorithm is presented in paper:
B. Kamiński, P. Szufel: Asynchronous Knowledge Gradient Policy for Ranking and Selection, Winter Simulation Conference, Proceedings of the 2014 Winter Simulation Conference, A. Tolk, S. Y. Diallo, I. O. Ryzhov, L. Yilmaz, S. Buckley, and J. A. Miller, eds., s. 3785-3796, 2014