Decision support systems apply the models of cooperative games with transferable utilities (TU games) for aggregating and ranking attributes based on their interdependent features. In TU games, players work together for a common goal under binding agreements. The contributions by the players to achieve this common goal generate some worth. A solution is a distribution of the total worth generated by all the players together. The Shapley value is the most popular solution of TU games. When cooperative game theoretic models are used in decision support systems, especially in Artificial Intelligence and Machine Learning models, the Shapley value is frequently used as a suitable solution, partly because it is easy to understand and partly because of its marginalistic approach. However, alternative criteria also exist in the literature that can be of interest to the practitioners. Most of the solutions based on these criteria differ by the property of a “non-productive player” whose contribution is marginal in different contexts of the game. We study how this property is used to characterize standard solution concepts of TU games. This, we believe will motivate the decision maker to explore possible alternative solutions that would treat the non-productive players differently from the Shapley assumptions.

决策支持系统应用具有可转移效用的合作博弈模型（TU 博弈），根据属性的相互依赖特征对属性进行聚合和排序。在 TU 游戏中，玩家在具有约束力的协议下为共同目标而共同努力。玩家为实现这一共同目标所做的贡献会产生一些价值。一个解决方案是分配所有玩家一起产生的总价值。Shapley 值是最流行的 TU 游戏解决方案。当合作博弈论模型用于决策支持系统时，尤其是在人工智能和机器学习模型中时，Shapley 值经常被用作合适的解决方案，部分原因是它易于理解，部分原因是其边缘化方法。然而，替代标准也存在于从业者可能感兴趣的文献中。大多数基于这些标准的解决方案因“非生产性玩家”的属性而异，其贡献在不同的游戏环境中是微不足道的。我们研究如何使用此属性来表征 TU 游戏的标准解决方案概念。我们相信，这将激励决策者探索可能的替代解决方案，以不同于 Shapley 假设的方式对待非生产性参与者。