Alternate distributed allocation of time reuse patterns in Fog-enabled cooperative D2D networks

Fog computing is regarded as a promising architecture to support various services on demand. Hand-held devices (or fog nodes) are becoming more and more powerful and can help each other to communicate and compute. In particular, the proximity-based communication through the device-to-device (D2D) links is one important paradigm of fog computing. In this paper, we look into the time reuse problem in a fog-enabled cooperative network. Even though the number of possible time reuse patterns increases exponentially with the quantity of the devices, we first show that the optimal network performance can be achieved by only activating a limited number of reuse patterns. Furthermore, when limited to a specific set of time reuse patterns, the optimal allocation problem becomes a biconvex one. Based on the alternate convex search (ACS) method and the alternating-direction method of multipliers (ADMM), a semi-distributed algorithm of low complexity, namely alternate distributed allocation of time reuse patterns (ADARP), is proposed to identify the optimal time reuse profile and the corresponding optimal association strategies. This semi-distributed nature renders our proposed solution suitable for fog-enabled cooperative D2D networks. Numerical simulations are carried out to corroborate our designs.