Harvesting Game: Strategic User Modeling, Mechanism Design and Equilibrium Analysis for Harvesting Resource
Track: Work in Progress
Categories: Defense, AI in Cybersecurity of Critical Systems, Reliability of Critical Systems, AI for detection and/or repair of service interruption
Keywords: Cloud Computing, Collusion, Game Theory, Mechanism Design
Abstract: Harvesting resources have emerged as a cutting-edge cloud computing solution, aimed at optimizing machine utilization and enhancing revenue for large-scale cloud platforms. By offering dynamically-sized virtual machines (VMs) at discounted prices, platforms can effectively address resource bottlenecks. However, existing pricing mechanisms are vulnerable to untruthful reporting and collusion, putting the platform's integrity and financial viability at risk.
To address this issue, we formalize the "Harvesting Game", a leader-follower game that incorporates strategies of lying and collusion among users and optimal mechanism design by the platform. Equilibrium analysis of Harvesting Game provides useful insights for both users and cloud platforms. Further experiments based on multi-agent learning under diverse settings are conducting.
Submission Number: 1
Loading