Neural Double Auction Mechanism

TL;DR

This paper introduces a neural network-based approach to automatically design double auction mechanisms that satisfy key economic properties, improving budget balance and efficiency over existing protocols.

Contribution

It presents a novel neural network architecture for automated mechanism design in double auctions, ensuring desirable economic properties and outperforming traditional protocols.

Findings

The learned mechanism is more budget-balanced than the VCG protocol.

It achieves higher economic efficiency than the MD protocol.

Incentive compatibility is mostly maintained.

Abstract

Mechanism design, a branch of economics, aims to design rules that can autonomously achieve desired outcomes in resource allocation and public decision making. The research on mechanism design using machine learning is called automated mechanism design or mechanism learning. In our research, we constructed a new network based on the existing method for single auctions and aimed to automatically design a mechanism by applying it to double auctions. In particular, we focused on the following four desirable properties for the mechanism: individual rationality, balanced budget, Pareto efficiency, and incentive compatibility. We conducted experiments assuming a small-scale double auction and clarified how deterministic the trade matching of the obtained mechanism is. We also confirmed how much the learnt mechanism satisfies the four properties compared to two representative protocols. As a result, we verified that the mechanism is more budget-balanced than the VCG protocol and more economically efficient than the MD protocol, with the incentive compatibility mostly guaranteed.