Runtime Verification of Multi-Agent Self-Adaptive System

Abstract

Self-adaptive system (SAS) used in the field needs to comply with strict quality requirements. In order to study the adaptation of multi-agent system to dynamic changes at runtime, an adaptive runtime verification method based on multi-agent system is proposed. The characteristic of this method is to convert the environmental impact factors in the multi-agent system into a probabilistic form, and perform runtime verification in a formal way to enable the system to regulate itself. The use of the method is discussed through a real case. The experimental results show that the method is suitable for an adaptive system with multiple agents.