Projects

DMASON a framework for Distributed and Parallel Agent-based Simulation


Agent-Based Models (ABMs) denote a class of models which, by simulating the behavior of multiple agents (i.e., independent actions, interactions and adaptation), aim to emulate and/or predict complex phenomena. One of the general features of ABM simulations is their experimental capacity, that requires a viable and reliable infrastructure to interact with a running simulation, monitoring its behaviour, as it proceeds, and applying changes to the configurations at run time, in order to study “what if” scenarios.

A common approach for improving the efficiency and the effectiveness of ABMs as a research tool is to distribute the overall computation on a number of ma- chines, which makes the design of the simulation model particularly challenging.

D-MASON framework is a distributed version of the MASON library for writing and running Agent-based simulations.

Summarizing, D-MASON provides:

  • Work partitioning, consists in the problem of decomposing a program to a set of processors (LPs) in order to achieve load balancing and efficient communication. D-MASON provides three different kinds of work partitioning: Uniform partitioning of the simulation field in which the simulation work load is divided equally among the processors, obviously this solutions does not consider that the agents can positioned in a non-uniform fashion over the filed; the second partitioning scheme consider that the agents can be placed and simulated according to a non-uniform distribution over the simulation field, and is named Non-uniform partitioning; the first two partitioning scheme exploits spatial relationships of the agents, so that use a space partitioning approach. The last partitioning scheme does not use the spatial relationships, instead, consider that the relationships between the agents are made using a graph. In such cases the simulations use a Network field for mapping the relationships between the agents. D-MASON provides a new distributed field, called DNetwork. This field is extremely useful in several ares such as biology (cellular networks), social and political science (communication and collaboration networks) as well as chemistry (metabolic network) and so on.
  • Memory consistency, in an ABM, the overall system evolves in discrete events (ideally all agents change their state simultaneously). However, the agents of a region are updated sequentially. In this case, the system, or the modeler, must ensure that the accesses to the states of the agents are consistent. D-MASON this problem is resolved, at framework-level, by exploiting the Java Method Handler mechanism.
  • Scalable communication,D-MASON LPs communicate via a well-known mechanism, based on the Publish/Subscribe (P/S) design pattern: a multicast channel is assigned to each cell; LPs then simply subscribe to the topics associated with the cells which overlap with their Area of Interest (AOI) in order to receive relevant message updates.D-MASON is designed to be used with any Message Oriented Middleware that implements the PS pattern, that enable to exploit heterogeneous computing power. Furthermore,D-MASON can also be deployed on HPC systems. In order to better exploit such homogeneous environments D-MASON provides also an MPI-based Publish/Subscribe communication.
  • Execution on Cloud Computing systems, SIMulation-as-a-Service (SIMaaS) infrastructure provides a very attractive prospective for the future environment to execute simulations, due to the good price-performance ratio.D-MASON provides easy-to-use system management based on Web technologies and tools to execute and visualize simulations on cloud computing systems.

Download and try it by yourself on GitHub.

 Android Sparkie


With Sparkie you can create wonderful word clouds directly on your smartphone!
Sparkie is easy, it doesn’t require internet connection, and it is totally free, without any advertisement. Choose you own text, build the cloud and share the most creative word clouds!
You can write any words, for example you have the possibility to load a song lyric by just adding artist name and song title or you can load a text from a web page by adding the website address.

     

Sparkie will analyze the text that you have choosen and will create an original word cloud. You could custom your clouds with free tools as color setting, fonts and orientation.
The word cloud built will show the most meaningful words in your text, in fact the words that are used more frequently in the text will appear bigger than the others. Use Sparkie to create word clouds to summarize and decorate business presentations, to abstract thoughts and quotes or just to create something unique and meaningful! With Sparkie is easy share your word clouds with your friends with Facebook, Google+, Twitter, Whatsapp, Instagram and other app, or you can save your creations in Sparkie’s gallery!
We will pleased to receive your feedback at weightwords@gmail.com or on Google+, to improve our work with your opinion and support!

Last Update 2017.
@ Carmine Spagnuolo.