Provenance and Traceability in Agent-based Policy Simulation

In the field of policy modelling, a trend to growing complexity of simulation models can be observed. One of the reasons for this development is the fact that for policy cases of practical interest there are usually no theories available from which "simple" simulation models could be derived. Instead a sometimes vast amount of information - positions and opinions of different interest groups or stakeholders, recorded e.g. in scenarios; background documents - has to be taken into account by the simulation model. Such models are usually referred to as "evidencebased models". In order to foster the use of the simulation method for decision-making, a constructive and structured approach for converting the evidence base into a simulation model is advised. This paper outlines such a novel approach, which has been developed within the OCOPOMO project. This approach basically consists of a policy development process specification, and a software toolbox supporting this process. Main focus of this paper are simulation related aspects of both the process and the toolbox, with the aim to point out importance and impact of maintaining provenance and traceability throughout the different stages of model development, simulation and analysis of simulation results.