Modeling and Simulating Adaptive Multi-Agent Systems with CAMLE

Abstract—With the advent of embedded and mobile computing techniques, software systems are increasingly operated in open and dynamic environments. Such systems desire self-adaptive capabilities. This paper proposes an agent-oriented approach to the modeling and simulation of distributed adaptive systems. The approach enables the designers to construct structural and behavioral models at a high abstraction level, and to validate the models at design stage through simulation. The modeling language supports devising decentralized adaption logic that enables various types of adaptions at both component and system levels. The simulation of adaptation processes can demonstrate how a system dynamically re-organizes in response to the changes in the operating context. Case studies show that our approach can support various self-adaptation mechanisms in which a multi-agent system adapts to internal failures, unexpected environment, or user's changing requirements