TY - JOUR
T1 - Stochastic game-based dynamic information delivery system for wireless cooperative networks
AU - Feng, Li
AU - Ali, Amjad
AU - Liaqat, Hannan Bin
AU - Iftikhar, Muhammad Aksam
AU - Bashir, Ali Kashif
AU - Pack, Sangheon
N1 - Funding Information:
This research was supported by National Research Foundation (NRF) of Korea Grant funded by the Korean Government (MSIP) (No. 2017R1E1A1A01073742 )
Publisher Copyright:
© 2019 Elsevier B.V.
PY - 2019/6
Y1 - 2019/6
N2 - The haptic communications is considered as the prime application running on the Tactile Internet. Therefore, Tactile Internet required to be highly reliable, provide a very low latencies, and required sufficient capacities at intermediate nodes to allow a large number of devices to communicate with each other simultaneously and autonomously. Moreover, the wireless cooperative network (WCN), is considered as one of the major component of the 5G technologies due to it promising advantages, such as improving wireless transmission capacity and reliability. However, the selfish nature of relay nodes may depress such enhancement and is not favored by the source node. In this paper, we propose an incentive-based dynamic flow allocation (FA) and forwarding strategy selection (FSS) scheme under time-varying selfishness. In the proposed scheme, the source node determines the FA to maximize the average network throughput under the constraints of network stability and selfishness boundaries, while each selfish relay executes the FSS to optimize its own profit with regard to the dynamic network state. Moreover, to cope with the conflicting interests between selfish relays a stochastic game model is employed to design a competition for haptic information forwarding and Nash equilibrium is proven also a combined Q-learning-based algorithm is proposed to guide the relays’ forwarding strategies. Furthermore, by considering the stochastic property of the network state, the FA for the source is formulated as a stochastic optimization problem. Finally, by exploiting the concept of virtual selfishness queue, the problem is solved by using the Lyapunov optimization theory. Performance of the proposed scheme is evaluated with traditional FA approach and data queue-based FA approach. Numerical results exhibit that our scheme not only sustains a large network throughput but also achieves low latency and avoids the occurrence of a completely selfish relay in the long term.
AB - The haptic communications is considered as the prime application running on the Tactile Internet. Therefore, Tactile Internet required to be highly reliable, provide a very low latencies, and required sufficient capacities at intermediate nodes to allow a large number of devices to communicate with each other simultaneously and autonomously. Moreover, the wireless cooperative network (WCN), is considered as one of the major component of the 5G technologies due to it promising advantages, such as improving wireless transmission capacity and reliability. However, the selfish nature of relay nodes may depress such enhancement and is not favored by the source node. In this paper, we propose an incentive-based dynamic flow allocation (FA) and forwarding strategy selection (FSS) scheme under time-varying selfishness. In the proposed scheme, the source node determines the FA to maximize the average network throughput under the constraints of network stability and selfishness boundaries, while each selfish relay executes the FSS to optimize its own profit with regard to the dynamic network state. Moreover, to cope with the conflicting interests between selfish relays a stochastic game model is employed to design a competition for haptic information forwarding and Nash equilibrium is proven also a combined Q-learning-based algorithm is proposed to guide the relays’ forwarding strategies. Furthermore, by considering the stochastic property of the network state, the FA for the source is formulated as a stochastic optimization problem. Finally, by exploiting the concept of virtual selfishness queue, the problem is solved by using the Lyapunov optimization theory. Performance of the proposed scheme is evaluated with traditional FA approach and data queue-based FA approach. Numerical results exhibit that our scheme not only sustains a large network throughput but also achieves low latency and avoids the occurrence of a completely selfish relay in the long term.
KW - 5G
KW - Incentive mechanism
KW - Stochastic game
KW - Tactile internet
KW - Virtual selfish queue
KW - Wireless resource allocation
UR - http://www.scopus.com/inward/record.url?scp=85059839212&partnerID=8YFLogxK
U2 - 10.1016/j.future.2019.01.011
DO - 10.1016/j.future.2019.01.011
M3 - Article
AN - SCOPUS:85059839212
SN - 0167-739X
VL - 95
SP - 277
EP - 291
JO - Future Generation Computer Systems
JF - Future Generation Computer Systems
ER -