2024 Volume 36 Issue 4 Pages 747-756
Understanding human flexible cooperative behavior in groups is crucial for developing intelligent systems that can collaborate effectively with humans. Group behavior involves top-down action selection based on shared intentions and bottom-up intention adjustments derived from individual behavior analysis. To construct agent models emulating this process, we analyzed the timing of dynamic intentional changes and strategies for adjusting individual intentions within the group. This analysis utilized a patterned task abstracting group behavior, including cooperative selection of intentions and actions to achieve specific goals. We constructed an agent model based on these findings and verified the action selection process through simulations.
