Adaptive collaboration while agents' performance changes periodically

An objective of Adaptive Collaboration (AC) is to maintain the performance of a group. This paper begins by clarifying AC. Next, the AC problem is formalized based on the Environment-Class, Agent, Role, Group, and Object (E-CARGO) model. Based on the AC problem model, we consider a typical scenario where individual agents periodically change their performance. To conduct AC, we propose an effective AC approach for such a scenario. Simulations are conducted to verify the benefits of AC. Results indicate that AC performs better than static collaboration in the assumed scenario and indicates a 15% performance increase on average with 3-5 role re-assignments in a period.