Department
Physics and Engineering
Location
Bethel University
Document Type
Poster
Start Date
2-25-2026 4:00 PM
End Date
2-25-2026 5:00 PM
Abstract
This paper examines traffic offloading between two wireless networks operating in the same service region, where a beneficiary network can temporarily redirect calls to a benefactor network during periods of congestion. To analyze this interaction, we develop a finite level‑dependent quasi‑birth‑and‑death (LD‑QBD) model that captures the joint dynamics of admission control, opportunistic offloading, and preemption. The LD‑QBD structure enables exact computation of the steady‑state distribution using efficient sparse linear solvers. From these steady‑state probabilities, we derive key performance metrics, including blocking and drop probabilities, mean occupancy, throughput, mean system time, admission gain, and successful throughput gain. Numerical results show that controlled offloading can substantially increase overall system capacity and improve user performance. The proposed modeling framework provides a robust analytical approach for evaluating heterogeneous wireless networks and emerging offloading strategies.
Recommended Citation
Tang, Shensheng, "Modeling and Analysis of Opportunistic Traffic Offloading between Wireless Networks" (2026). Wednesday, February 25, 2026. 13.
https://spark.bethel.edu/dayofscholarship/spring2026/spr2026/13
Included in
Modeling and Analysis of Opportunistic Traffic Offloading between Wireless Networks
Bethel University
This paper examines traffic offloading between two wireless networks operating in the same service region, where a beneficiary network can temporarily redirect calls to a benefactor network during periods of congestion. To analyze this interaction, we develop a finite level‑dependent quasi‑birth‑and‑death (LD‑QBD) model that captures the joint dynamics of admission control, opportunistic offloading, and preemption. The LD‑QBD structure enables exact computation of the steady‑state distribution using efficient sparse linear solvers. From these steady‑state probabilities, we derive key performance metrics, including blocking and drop probabilities, mean occupancy, throughput, mean system time, admission gain, and successful throughput gain. Numerical results show that controlled offloading can substantially increase overall system capacity and improve user performance. The proposed modeling framework provides a robust analytical approach for evaluating heterogeneous wireless networks and emerging offloading strategies.