Local-Global Agent Failover Based on Organizational Models

This contribution introduces a failover strategy for agent lookup in workflows. It is based on an organizational model and relations between different model elements. The algorithms that evaluate the organizational model prioritize local before global knowledge. Local knowledge is better suited to handle exceptions than global rules. The algorithms are able to find the correct agents even in unplanned situations. Starting point for the algorithms is a description of the requested agents in a language expression. This description is based on the location of the agent(s) within the organizational structure, or on different relations that exist within the model.