Multi-agent systems are widely used in implementing systems with autonomous behaviour, as well as simulating models that involve human behaviour. Ethical decision-making is a unique aspect of such socio-cognitive systems. In situations that require deliberation over a multitude of options with ethical implications, moral dilemmas are often resolved by a range of heuristics and principles that view the situation from different perspectives. While constraint and utility-driven decision-making strategies are customary in agent-based models, the incompatibility among multiple perspectives can invalidate the underlying premises of models of probabilistic utility-based decisions that rely on classic Kolmogorov axioms. In this paper, we highlight the fallacies of such models and present a practical case study to demonstrate how quantum cognition models can provide an alternative and credible representation of decision-making behaviour in simulations that involve ethical dilemmas.

多主体系统广泛用于实现具有自主行为的系统，以及模拟涉及人类行为的模型。道德决策是这种社会认知系统的独特方面。在需要考虑具有伦理含义的多种选择的情况下，道德困境通常由一系列启发式方法和原则解决，这些启发式方法和原理从不同的角度审视情况。虽然约束和效用驱动的决策策略在基于代理的模型中很常见，但多种观点之间的不兼容会使基于概率效用的决策（依赖经典Kolmogorov公理）模型的基础前提无效。在本文中，