The ethical shortlisting problem
研究如何用线性规划将功利主义、最大最小、平等主义等抽象道德理论应用于招聘初选，为每个求职者计算加权评分并排序，实验表明模型可扩展且结果可解释，推荐使用最大最小模型逐步淘汰最低分者。
Hiring is a fundamental, frequent activity for all organizations. Hiring decisions have been reported to be subject to conscious and unconscious biases in the literature. The field of Computational Ethics aims to quantify and maximize the ethicality of decisions. This paper attempts to apply Computational Ethics to the shortlisting process in hiring through the use of Linear Programming. Given a set of applicants for a job with numerical qualification values, the author aims to determine weights for each qualification type to compute scores and resulting rankings for each applicant. To this end, Abstract Moral Theories of Utilitarianism, Maximin/Leximin, Egalitarianism, and Prioritarianism are utilized and applied to a set of randomly generated applicant data. Computational experiments demonstrate that the models are scalable and return interpretable results. The necessity of a quota-based shortlisting system to alleviate disadvantaged candidates is highlighted. The author recommends the use of the Maximin model and iteratively eliminating the applicant with the lowest score.