TY - GEN
T1 - Cost and latency tradeoff in mobile edge computing: A distributed game approach
AU - Zaw, Chit Wutyee
AU - Ei, Nway Nway
AU - Han, Yeo Reum Im
AU - Tun, Yan Kyaw
AU - Hong, Choong Seon
PY - 2019/2/27
Y1 - 2019/2/27
N2 - Resource allocation has been a critical issue in Mobile Edge Computing (MEC) since MEC servers colocated at Base Stations (BS) are limited in resources and users are competitive in nature. Users' devices have restraint on power and computation capability. Moreover, users want to offload their tasks to meet their latency deadlines and they have cost for offloading their tasks. In this paper, we analyze this tradeoff between cost and latency for offloading tasks. Basically, if users want to have less latency, they need more resources to accomplish this. But, more resources to allocate to users means they have to pay more cost. In addition, we formulate the resource allocation as Generalized Nash Equilibrium Problem because users' strategies are conflicted with one another when they compete for the resources from the resources pool at the BS. We propose a distributed algorithm for the game formulation since users prefer to control their own resources rather than revealing their information to others. Then, we analyze the Price of Anarchy numerically.
AB - Resource allocation has been a critical issue in Mobile Edge Computing (MEC) since MEC servers colocated at Base Stations (BS) are limited in resources and users are competitive in nature. Users' devices have restraint on power and computation capability. Moreover, users want to offload their tasks to meet their latency deadlines and they have cost for offloading their tasks. In this paper, we analyze this tradeoff between cost and latency for offloading tasks. Basically, if users want to have less latency, they need more resources to accomplish this. But, more resources to allocate to users means they have to pay more cost. In addition, we formulate the resource allocation as Generalized Nash Equilibrium Problem because users' strategies are conflicted with one another when they compete for the resources from the resources pool at the BS. We propose a distributed algorithm for the game formulation since users prefer to control their own resources rather than revealing their information to others. Then, we analyze the Price of Anarchy numerically.
UR - https://ieeexplore.ieee.org/abstract/document/8679304
U2 - 10.1109/BIGCOMP.2019.8679304
DO - 10.1109/BIGCOMP.2019.8679304
M3 - Article in proceeding
BT - In Proc. IEEE International Conference on Big Data and Smart Computing (BigComp)
ER -