Modeling and Simulating Adaptive Multi-agent Systems with CAMLE

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 to construct structural and behavioral models at a high abstraction level, and to validate the models at design stage through model simulation. The modeling language facilitates devising decentralized adaption logic at both component and system levels. Simulation demonstrates 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 with which a multi-agent system adjusts to internal failures, unexpected environment, or user´s changing requirements.