Simulation-based design and evaluation of multi-agent systems

Multi-agent systems (MAS) provide powerful modelling concepts for representing real-world applications with an appropriate degree of complexity and dynamics [1]. Several academic and industrial experiences have already shown that the use of MAS offers advantages in many different areas such as manufacturing processes, e-Commerce, and network management. Since MAS in such contexts need to be tested before their deployment and execution in real operating environment, methodologies that support system validation through simulation (e.g. discrete-event simulation, agent-based simulation, etc.) are highly required. In fact, simulation of MAS cannot only demonstrate that MAS correctly behaves according to its specifications but can also support the analysis of emergent properties of the MAS under-test [2,3]. In the last few years there has been an increasing number of initiatives to develop agent-based methodologies for the development of software agent-based systems (e.g. GAIA, Tropos, Prometheus, INGENIAS, PASSI, ADELFE, PASSIM) and simulation tools for the analysis of complex systems modelled as multi-agent systems (Repast, Repast Simphony, Swarm, Netlogo, Mason). As ‘‘Agent-oriented Software Engineering” (AOSE) [4] and ‘‘Agent-based Modelling and Simulation” (ABMS) [5] are recognized as very interesting emerging paradigms that will have a major impact on the quality of science and society over the next years, their knowledge will help position the researchers and practitioners at the forefront of the field. Therefore, we believe that the methodological and technological trends in the convergence and integration of these areas need to be widely explored to provide an exclusive research roadmap to both ABMS and AOSE communities. The aim of this special issue is to provide a comprehensive guide on new ideas and results in the integration of Simulation of Multi-Agent Systems and Agent-oriented Software Engineering domains. It captures the state-of-the-art in such domains in terms of techniques and methodologies for agent-based modelling and simulation, simulation-driven development processes for multi-agent systems, and simulation-oriented analysis of emergent agent behaviours in complex multi-agent systems. It also identifies potential research directions and technologies that will drive innovations within this domain. We expect the papers of this special issue to serve as a valuable reference for larger audience such as software architects, practitioners, developers, researchers, and students. The nine papers of this special issue successfully capture the state of the art in the integrated AOSE and ABMS domains, and offer very interesting perspectives both on current trends and on the innovation forces that will drive research and development efforts in the coming years. In particular, significant key issues are tackled in the papers, among which: model driven development (MDD) based methodologies for agent-based simulation, modelling collaborative agent behaviours in multi-agent systems, agent-based techniques to solve real-time scheduling problems in complex systems, self-organization as key element to drive adaptive system behaviour, analysis of interaction in reputation and trading systems, etc. The paper by José Alberto Araúzo, Javier Pajares, Adolfo Lopez-Paredes, ‘‘Simulating the dynamic scheduling of project portfolios”, shows how agent-based modelling can help to solve the problem of dynamic scheduling of resources for multiple concurrent projects in real time. Mathematical approaches, like integer programming or network based techniques, cannot describe the complexity of multiple projects environments, which have many interrelated elements, and specially to adapt the analysis to dynamic changes. MAS allow capturing real complexity, and managing the dynamical issues of the environment. In this case, projects compete for resources, whose price evolves in real time depending on current demand. As a result of the negotiation among project manager agents and resource manager agents, the prices of resources are established dynamically, changing in time according to actual demand, and the cost of a project is also calculated dynamically, taking into account the others and resources availability and cost. This ability to adapt to a changing environment, and the way prices emerge from interaction among projects and resources, are illustrative of the potential of agent-based simulation for the analysis of complex systems. The use of agent-based simulation for open markets analysis is also illustrated by the work of Paulo Trigo, Paulo Marques, Helder Coelho, ‘‘Virtual Agents for Running Electricity Markets”. Here, the focus is on deregulated energy markets, where competition should be achieved in a fair and transparent way. The implementation of an artificial market allows exploring whether the application of market rules drive to a coherent market behaviour, which emerges from the overall simulated environment. This can be validated in different conditions, where there are companies of different sizes and market share,