Computational offloading with delay and capacity constraints in mobile edge

Due to the limited computing resources in mobile devices, it is desirable to offload the computation intensive and time sensitive tasks to the network edge. This paper aims at optimizing the total consumption cost incurred by the usage of the limited computational resources in mobile edge (e.g. small cell, access point, macro base station) while satisfying all the computational demands from each network edge node within a certain time delay. We model this problem as an Integer Programming problem and then propose a Two-Phase Optimization (TPO) algorithm and an Iterative Improvement (Π) algorithm for the purpose of optimality and computational efficiency respectively. The experimental results demonstrate that our proposed TPO finds an optimal solution up to 10 times faster than directly solving the minimization problem using a standard optimizer, while the II algorithm could strike a good tradeoff between the solution quality and the running time incurred.