Multi-Agents Systems (MAS) are modern approaches that need an additional investigation to improve their reliability and adaptability levels. Exception management is one way to reach this goal and this paper is dedicated to this specific subject. The purpose of this document is to examine the exception concept in MAS domain and to suggest a model adjusted to MAS challenges such as heterogeneity, openness and particularly agents' autonomy. Previous attempts in the agent's society have concluded set of findings that demonstrated the necessity of exception handling in MAS at the system level. The handling includes management and the needed processes related to management. The attainment up to now can be applied only to special MAS type. Usually, agents are nonautonomous and the system-level strategies need an impeccable cooperation between agents in the exception handling process. In our proposed model, the agent's ability to approach exceptions by itself is considered as a prerequisite to assure agent autonomy. Then, exception handling depends on agent-level processes to deal with the limitations of contemporary attainments and thus, they are complementary. Agent preserves the ability to independently decide when to activate exception handling and when to receive system-level help or believe in its skills.
CITATION STYLE
Athamena, B., & Houhamdi, Z. (2017). An exception management model in multi-agents systems. Journal of Computer Science, 13(5), 140–152. https://doi.org/10.3844/jcssp.2017.140.152
Mendeley helps you to discover research relevant for your work.