Complete Decentralized Mechanism Design for Online Machine Scheduling

We study an online version of the classical parallel scheduling problem to minimize the total weighted completion time in a total decentralized setting: both the jobs and the machines are selfish entities who try to maximize their own interest. We study the setting from an algorithmic mechanism design and present a polynomial time decentralized scheduling mechanism that induces the rational jobs and machines to report truthfully about their private information, and lures the jobs to choose proper machines such that the resulting schedule is 3.281-competitive.