Parameters.

Organisations rely upon group formation to solve complex tasks, and groups often adapt to the demands of the task they face by changing their composition periodically. Previous research has often employed experimental, survey-based, and fieldwork methods to study the effects of group adaptation on task performance. This paper, by contrast, employs an agent-based approach to study these effects. There are three reasons why we do so. First, agent-based modelling and simulation allows to take into account further factors that might moderate the relationship between group adaptation and task performance, such as individual learning and task complexity. Second, such an approach allows to study large variations in the variables of interest, which contributes to the generalisation of our results. Finally, by employing an agent-based approach, we are able to study the longitudinal effects of group adaptation on task performance. Longitudinal analyses are often missing in prior related research. Our results indicate that reorganising well-performing groups might be beneficial, but only if individual learning is restricted. However, there are also cases in which group adaptation might unfold adverse effects. We provide extensive analyses that shed additional light on and help explain the ambiguous results of previous research.

组织依托群体组建以解决复杂任务，而群体通常会通过周期性调整成员构成，适配其所面临的任务需求。过往研究多采用实验法、问卷调查法与实地调研方法，探究群体适配对任务绩效的影响。与之相对，本文采用基于智能体（Agent-based）的研究方法开展此类研究。本文采用该方法主要基于三点动因：其一，基于智能体的建模与仿真能够纳入更多可调节群体适配与任务绩效间关联的因素，例如个体学习与任务复杂度；其二，该方法可对目标变量开展大范围变量变化的研究，有助于提升研究结果的可推广性；其三，采用基于智能体的研究方法，我们能够探究群体适配对任务绩效的长期影响，而这类纵向分析在既往相关研究中往往较为匮乏。研究结果表明，重组绩效优异的群体或可带来益处，但该结论仅在个体学习受到限制的前提下成立。不过，部分场景下群体适配也可能产生负面效应。本文通过大量分析，进一步阐明并解释了既往研究中出现的歧义性结果。