Qualitative Modeling of MAS Dynamics Using Systemic Modeling to Examine the Intended and Unintended Consequences of Agent Coaction

摘要

The design of agent-based software applications is supported by modeling approaches that focus on the design of individual agents as well as their arrangement in organizational structures. However, it is a challenging task to deduce the collective system behavior from the sum of designs of the autonomous, pro-active actors and their collaboration can lead to surprising effects. Therefore, developers of agent collectives require tools to plan for the effects of agent coaction. Here, we propose a systemic modeling level that supplements established agent-oriented modeling approaches and utilizes System Science modeling concepts to support qualitative examinations of the system-wide dynamics that can result from agent coaction. We discuss the systematic derivation of systemic MAS abstractions from established design notations and exemplify systemic modeling of MAS in a case study where the interdependencies of agent activites and their consequences are identified at design-time and compared to run-time effects.