Abstract
We have conducted research on the feasibility of a system for sharing crop-growth data distributed on a network. We have shown that some farmers' organizations have potential to evolve into platforms for sharing crop-growth data. From a software development viewpoint, we have developed an experimental cooperative agent-based system to investigate its implementation requirements. Our experimental system consists of 18 virtual servers, and 751 soybean growth records distributed among these virtual servers. Our cooperative agents have successfully exchanged and shared these data.
When we try to implement systems for sharing distributed databases, heterogeneities between databases often cause conflicts. To resolve such heterogeneities, we utilized a cooperative agent system and ontology in our experimental system. In a network of cooperative agents system, agents have to communicate with each other in a standardized manner. In our experimental system, we adopted KQML (Knowledge Query and Manipulation Language) as our standard.•@KQML agents communicate with each other by exchanging messages. The messages describe both directives (e.g. query, answer, notification) and contents. When receiving a KQML message, an agent processes the contents of the message according to the directive. To resolve semantic conflicts, agents refer to ontologies, and resolve conflicts. In our ontology, a hierarchy of categories and variations in notation are described.
