5 juillet 2020
Ce document est lié à :
info:eu-repo/semantics/altIdentifier/arxiv/1903.10361v3
Ce document est lié à :
info:eu-repo/grantAgreement//740435/EU/MDDS (Mechanism Design for Data Science)/ERC-2016-ADG
Anna Bogomolnaia et al., « On the fair division of a random object », HAL-SHS : économie et finance, ID : 10670/1.mns7kv
Ann likes oranges and dislikes apples; Bob likes apples and dislikes oranges. Tomorrow theywill receive one fruit that will be an orange or an apple with equal probability 0.5. Giving toeach half of that fruit is fair for each realisation of the fruit; but agreeing that whatever fruitappears will go to the agent who likes it more gives a higher expected utility to each agent andis fair in the average sense: in expectation, each agent prefers his allocation to the equal divisionof the object, he gets a “Fair Share”.We turn this familiar observation into an economic design problem: upon drawing a randomobject (the fruit), we learn the realised utility of each agent and can compare it to the mean of hisdistribution of utilities; no other statistical information about the distribution is available. Wefully characterize the division rules that use only this sparse information in the most efficientpossible way, while giving everyone a Fair Share. Although the probability distribution ofindividual utilities is arbitrary and mostly unknown to the designer, these rules perform in thesame range as the best fair rule having full knowledge of this distribution.