Abstract:
Multi-access edge computing (MEC) is emerging as
a solution to serve offloaded tasks from mobile devices that are
computing intensive and have very low latency and high
bandwidth requirements. Since compute resources are limited at
the MEC hosts, collaboration among hosts could enhance the
capabilities of sharing limited resources while minimizing the cost
of such hosts. However, the selection of optimal hosts to instantiate
the user applications is a major challenge when considering the
total service provisioning cost. In this paper, we formulate the
MEC hosts selection problem as a binary integer problem with the
objective to minimize the total cost of providing the offloading
services. We extend the Balas-Geoffrion algorithm to solve the
special case of binary programming problems similar to MEC host
selection problem. The time complexity of the MEC host selection
problem is therefore minimized. We show that our modified
algorithm outperforms Balas-Geoffrion algorithm in the number
of iterations required to reach the optimal solution. Then we
conduct an extensive simulation to show that the overall quality of-service of the MEC system is improved by the MEC hosts
collaborations in a limited bandwidth scenario by up to 13%.
However, the tradeoff is an increase in the cost of provisioning the
services.
