JAS 2: a new Java platform for AB modeling and dynamic microsimulation.

We present JAS 2, a new Java platform which aims at providing a unique simulation tool for discrete-event simulations, including agent-based and dynamic microsimulation models. JAS 2 is not released as a self-contained stand-alone application for model development. With the aim to develop large-scale, computationally resource-intensive models, the main architectural choice of JAS 2 is to use whenever possible standard, open-source tools aready available in the software development community. This decision entails a longer learning curve for the user than a dedicated, language-specific software package, but it allows for a more professional approach to simulation modeling compared to specific proprietary environments. There are two main advantages of using an open architecture where external libraries can be added and utilized: one is the possibility to easily integrate external functions that can be added as plug-ins; the other is the possibility to freely extend and modify the functions available on the platform, possibly within a cooperative effort of the community of users. Moreover, an open source application makes it easier to share and test the models. Building on the vast number of software solutions available, JAS 2 allows the user to separate data representation and management from the implementation of processes and behavioral algorithms. It also allows the user, within an object-oriented paradigm, to model the real-life system naturally and intuitively: each category of individuals or objects that populate the model is represented by a specific class, with its own variables and methods. The main value added of the platform lies in the potential for integration with RDBMS (relational database management) tools through ad-hoc microsimulation Java libraries. The management of input data persistence layers and simulation results is performed using standard database management tools, and the platform takes care of the automatic translation of the relational model (which is typical of a database) into the object-oriented simulation model.