主 題：Balancing Efficiency and Equality in Vehicle Licenses Allocation (兼顧公平和效率的車牌分配機(jī)制)
主講人：陳洲 博士 上海財(cái)經(jīng)大學(xué)信息管理與工程學(xué)院
主持人：陳夢(mèng)妙 博士 上海對(duì)外經(jīng)貿(mào)大學(xué)國(guó)際經(jīng)貿(mào)學(xué)院物流管理系
時(shí) 間：2021年10月28日（星期四）10:00—11:30
地 點(diǎn)：上海對(duì)外經(jīng)貿(mào)大學(xué) 博識(shí)樓113
主講人簡(jiǎn)介：陳洲，香港科技大學(xué)博士，上海財(cái)經(jīng)大學(xué)和香港科技大學(xué)博士后。研究方向包括運(yùn)籌學(xué)，博弈論，算法設(shè)計(jì)及其在網(wǎng)絡(luò)經(jīng)濟(jì)，收益管理，市場(chǎng)運(yùn)營(yíng)，共享經(jīng)濟(jì)中的應(yīng)用。在Discrete Applied Mathematics，Journal of Combinatorial Optimization和IEEE Transactions on Cloud Computing（SCI 一區(qū)）等國(guó)際權(quán)威雜志上發(fā)表論文多篇。
摘要：在目前汽車尾氣帶來(lái)的環(huán)境問(wèn)題和城市交通日益擁堵的大環(huán)境下，越來(lái)越多的國(guó)家和地區(qū)已經(jīng)開(kāi)始采取限制車牌數(shù)量的政策。在這些地方，政府對(duì)車牌進(jìn)行限額分配。目前，車牌的分配因地而異，已經(jīng)實(shí)施的機(jī)制有拍賣、搖號(hào)、保留價(jià)格搖號(hào)以及同時(shí)拍賣和搖號(hào)，不同的機(jī)制對(duì)公平和效率有不同的側(cè)重。我們希望設(shè)計(jì)一個(gè)最佳的、簡(jiǎn)單的機(jī)制來(lái)最優(yōu)地平衡效率和公平。我們首先提出了一個(gè)兩組的統(tǒng)一框架，該框架在實(shí)踐中包括或優(yōu)于所有現(xiàn)有的機(jī)制。在這個(gè)框架下，我們證明了只要參與者的價(jià)值分布是任一常用分布，最優(yōu)機(jī)制總是先拍賣后搖號(hào)。這種最佳的分配規(guī)則僅取決于車牌總數(shù)和參與者總數(shù)。最后我們將二組的框架擴(kuò)展為通用的多組框架，并且證明了兩組的最優(yōu)已經(jīng)接近理論的多組最優(yōu)。并且兩組最優(yōu)由于其激勵(lì)相容性和結(jié)構(gòu)的簡(jiǎn)單，在實(shí)踐中也易于實(shí)施。我們的工作為車牌分配問(wèn)題提供了一種簡(jiǎn)單易行且有效的工具，并且保證了在一定公平水平的約束下最大限度地提高社會(huì)效率。
Abstract: Many big cities have begun to adopt policies limiting the number of vehicle licenses because of traffic and air quality concerns. In these cities, every one or two months a limited number of new vehicle licenses are distributed amongst a very large number of potential car buyers. Currently, the allocation mechanisms differ from city to city. Several mechanisms have been developed and implemented, such as auctions, lotteries, lotteries with reserved price, and simultaneous auctions and lotteries. In this work, we attempt to design an optimal, simple mechanism to balance efficiency and equality. We first propose a framework using a unified two-group mechanism that either includes or outperforms all existing mechanisms in practice. Further, the unified framework is also easy to implement in practice due to its truthfulness and simple structure. Using this framework, assuming the players' private values are drawn independently from a common distribution, we prove the optimal mechanism always consists of a sequential auction and lottery. In addition, the optimal allocation rule depends only on the total number of players and the total number of licenses for all commonly used distributions. We then extend the two-group framework to a general multi-group framework. The experimental results show us that the optimal two-group mechanism is the best practical choice. Consequently, our work provides an effective tool for social planners to design truthful mechanisms to maximize social efficiency for any level of equality. We also discuss possible applications of our results to resource allocation in other settings.