Market Design for Task Offloading under Information Asymmetry and System Uncertainty

We design a market for task offloading in mobile edge computing under information asymmetry and unknown and uncertain system information. There exist multiple mobile devices (MDs), multiple edge servers (ESs), and a broker in the market. The challenges of the market design are two-fold. The MDs’ utilities and the ESs’ costs are their private information and unknown to the broker. Additionally, the MDs and ESs have no prior knowledge of system information to assist task offloading. Although the existing incentive mechanisms can help elicit the private information, the correlation between the private information and system information has not been formally studied. To fill the gap, we formulate a social welfare maximization problem with the constraint of total delay and decompose it into two sub-problems: ES selection problem and task allocation problem. Furthermore, we employ the multi-armed bandit theory to develop the online learning-based iterative double auction algorithm (OL-IDA), which can solve the above two sub-problems. We present a theoretical analysis to evaluate the performance of this algorithm in terms of the regret, violation and the desirable economic properties. Extensive simulation results demonstrate that the OL-IDA algorithm significantly maximizes the social welfare while guaranteeing the constraint of total delay.