Designing Efficient Market Mechanisms with Simplicity Requirements

点击翻译按钮获取中文摘要

This research proposal aims to investigate the notion of simplicity in the design of market mechanisms. The main motivation for this study is the observation that simple and transparent mechanisms are more likely to be used in practice due to the ease of understanding the mechanism's properties from the point of view of the participants (i.e., traders in the market). Simple mechanisms moreover tend to be easy to implement, and generally are not time-consuming to run. The term "market mechanism" refers to the platform and protocols by which a market operates. A market is a two-sided platform, where both buyers and sellers are participating agents, and both types of agents follow a trading mechanism set out by an intermediating party. The proposed research is of a theoretical nature, yet is motivated by practical applications. One of the main practical motivators is the ever-rising presence of e-commerce, and the need for transparent and efficient market mechanisms in e-commerce applications. The challenges addressed in this proposal are relevant to a range of large-scale applications, such as those in the sharing/access economy (where ride-sharing platforms and lodging platforms form well-known examples; they are applications that function essentially as market mechanisms), online marketplaces, and online advertising mechanisms. The need for simplicity in mechanism design has furthermore been called for in various recent works in micro-economic theory and computer science.The research proposal captures market platforms in an abstract framework into which the above application domains can be cast. It then plans to investigate through two work packages which desirable properties can possibly be guaranteed by which market mechanisms, while achieving a high performance both economically (i.e., the resulting social welfare should be high) and computationally (i.e., the runtime of such a mechanism should be low). The properties we require from these mechanisms are the economically standard notions of strategy-proofness, individual rationality, and budget-balance. Furthermore, we require our mechanisms to be "simple" and will investigate suitable formalisations of this concept. The two work packages examine the problem into two opposing directions: The first work package intends to analyse the most atomic case within our framework, where only one buyer, one seller, and one item is present. The second work package aims to design mechanisms for increasingly more general instances of the framework (e.g., considering many buyers and sellers, and many items with increasingly complex and possibly interdependent valuations for these items among the buyers and sellers).