Introduction

While many and heterogeneous models exist nowadays to assess the impact of government policies or market rules over agricultural systems, very few (if any) address the basic concerns of the academic community of transparency and reproducibility of experiments. Instead, most times the reader must be satisfied with validation experiments conducted by the researchers themselves.

To address this needs, we chose to develop our own model as Free Software, that is, its source code is publicly available and can be freely used and modified.

RegMAS (Regional Multi Agent Simulator) is a spatially explicit multi-agent model framework, developed in C++ language designed for long-term simulations of effects of government policies on agricultural systems (farm sizes, incomes, land use..).

More specifically, RegMAS conceives rural social systems (and in particular agricultural ones) as complex evolving systems, made of an heterogeneous set of individual acting ``agents'' (that is, farmers).

The researcher typically set the initial conditions where the farmers operate and ``observe'' their behaviours in such environment. Initial conditions can include pre-defined shocks along the temporal dimension, typically used to model changes in the policy or macro-economic variables (e.g. decoupling options, activities gross margins..).

This document will help you discover RegMAS and hopefully will make your life easier while using it for your own simulation exercises.

It is structured as follow.

Section 2 briefly describes the history and the methodology behind RegMAS, referring for details on several papers that discuss multi-agent models applied on agricultural systems.

Section 3 will guides users from obtaining the program with associated sample data trough having it working on their pc. It covers a typical MS Windows installation, while the compilation of RegMAS directly from the source code is covered in a corresponding section on its web-site wiki.

Once the program is installed, sections 4 to 7 will gradually introduces in the usage of RegMAS. Section 4 presents the program user's interface and the input files, allowing to run a complete RegMAS experiment using the predefined scenarios over the sample data. Users interested in creating their own scenario should consult section 5. A further point that most RegMAS users will likely be interested is how to apply the model over their own case-study region. This requires obtaining farm-specific and regional-level data (including spatial one) and it is covered on section 6.

Finally some researcher may want to personalise RegMAS, maybe changing the farmers behaviours or adding new details to the output. As previously noted, source code of RegMAS is fully available and users are encouraged to modify it for their own modelling needs. Section 7 gives a brief tour of the RegMAS Application Programming Interface (API), presents the tools used to develop RegMAS and point on the specific place where to get more help.

Once a user has became familiar on creating new scenarios or models, it will be time to analyse the results. Sections 8 and 9 respectively describe the kind of results that are saved by RegMAS and how to export them into a GIS for further analysis (the latter task is optional but recommended).

This manual ends with an Appendix covering the sources where further help in using RegMAS may be available and the license terms.

Have fun with RegMAS !