Mean-field-type game-based computation offloading in multi-access edge computing networks

Reginald A. Banez, Hamidou Tembine, Lixin Li, Chungang Yang, Lingyang Song, Zhu Han, H. Vincent Poor

Research output: Contribution to journalArticlepeer-review


Multi-access edge computing (MEC) has been proposed to reduce latency inherent in traditional cloud computing. One of the services offered in an MEC network (MECN) is computation offloading in which computing nodes, with limited capabilities and performance, can offload computation-intensive tasks to other computing nodes in the network. Recently, mean-field-type game (MFTG) has been applied in engineering applications in which the number of decision makers is finite and where a decision maker can be distinguishable from other decision makers and have a non-negligible effect on the total utility of the network. Since MECNs are implemented through finite number of computing nodes and the computing capability of a computing node can affect the state (i.e., the number of computation tasks) of the network, we propose non-cooperative and cooperative MFTG approaches to formulate computation offloading problems. In these scenarios, the goal of each computing node is to offload a portion of the aggregate computation tasks from the network that minimizes a specific cost. Then, we utilize a direct approach to calculate the optimal solution of these MFTG problems that minimizes the corresponding cost. Finally, we conclude the paper with simulations to show the significance of the approach.

Original languageEnglish (US)
Article number9195764
Pages (from-to)8366-8381
Number of pages16
JournalIEEE Transactions on Wireless Communications
Issue number12
StatePublished - Dec 2020


  • Mean-field-type games
  • computation offloading
  • multi-access edge computing networks

ASJC Scopus subject areas

  • Computer Science Applications
  • Electrical and Electronic Engineering
  • Applied Mathematics


Dive into the research topics of 'Mean-field-type game-based computation offloading in multi-access edge computing networks'. Together they form a unique fingerprint.

Cite this