Skip to Main Content
Multi-agent systems (MAS) are efficient solutions for commercial applications such as information retrieval and search. In a MAS, agents are usually designed with distribution of functionality and control. Lack of central control implies that the quality of service of MAS may be degraded because of possible unwanted behavior at the runtime, commonly known as emergent behavior. Detecting and removing emergent behavior during the design phase of MAS will lead to huge savings in deployment costs of such systems. An effective approach for the MAS design is to describe system requirements using scenarios. A scenario, commonly known as a message sequence chart or a sequence diagram, is a temporal sequence of messages sent between agents. In this paper a method for detecting emergent behavior of MAS by detecting incompleteness and partial description of scenarios is proposed. The method is explained along with a prototype MAS for semantic search that blends the search and ontological concept learning.