Delay-Tolerant Constrained OCO with Application to Network Resource Allocation

Wang, Juncheng; Liang, Ben; Dong, Min; Boudreau, Gary; Abou-zeid, Hatem

doi:10.1109/infocom42981.2021.9488698

Cited by 7 publications

(9 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, long-term constraints arise in many practical applications such as energy control in wireless communications, queueing stability in cloud networks, and power balancing in smart grids. Existing algorithms for OCO with long-term constraints can be categorized into saddle-point-typed algorithms [17]- [20] and virtual-queue-based algorithms [10], [21]- [23].…”

Section: B Oco With Long-term Constraintsmentioning

confidence: 99%

“…In [22], the virtual-queue-based algorithm was further extended to provide a dynamic regret bound. The impact of multi-slot feedback delay on constrained OCO was considered in [23] with dynamic regret analysis.…”

Section: B Oco With Long-term Constraintsmentioning

confidence: 99%

“…However, all of the above works on constrained OCO [10], [17]- [23] are under the standard per-time-slot update setting. No other known work considers periodic updates for OCO.…”

Section: B Oco With Long-term Constraintsmentioning

confidence: 99%

“…For the case of per-time-slot updates, the dynamic benchmark was originally proposed for OCO with short-term constraints [7] and was modified in [20], [22], [23] to incorporate long-term constraints. Here, we generalize it to account for periodic updates.…”

Section: B Regrets and Constraint Violationmentioning

confidence: 99%

“…Note that the constraint violation defined in [10], [17]- [23] are under the standard per-time-slot update setting. In contrast, our model accommodates the possibly time-varying update periods of multiple time slots.…”

Section: B Regrets and Constraint Violationmentioning

confidence: 99%

See 4 more Smart Citations

Periodic Updates for Constrained OCO with Application to Large-Scale Multi-Antenna Systems

Wang,

Dong,

Liang

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

In many dynamic systems, such as wireless communications, decisions on system operation are updated over time, and the decision maker requires an online learning approach to optimize its strategy in response to the changing environment. When the loss and constraint functions are convex, this belongs to the general family of online convex optimization (OCO). In existing OCO works, the environment is assumed to vary in a time-slotted fashion, while the decisions are updated at each time slot. This model is inadequate for systems that permit only periodic decision updates, i.e., each decision is fixed over multiple time slots, while the environment changes between the decision epochs. In this work, we consider periodic decision updates for OCO. We aim to minimize the accumulation of time-varying convex loss functions, subject to both short-term and long-term constraints. Information about the loss functions within the current update period may be incomplete and is revealed to the decision maker only after the decision is made. We propose an efficient algorithm, termed Periodic Queueing and Gradient Aggregation (PQGA), which employs novel periodic queues together with aggregated gradient descent to update the decisions over time. PQGA is applicable to both constant and time-varying update periods.Most importantly, we show that PQGA yields bounded dynamic regret, static regret, and constraint violation. Furthermore, they are sublinear over time if the accumulated variation of the system states and update periods do not grow too fast. As an example application, we study the performance of PQGA in a large-scale multi-antenna system shared by multiple wireless service providers. Simulation results show that PQGA converges fast and substantially outperforms the known best alternative.

show abstract

Section: B Oco With Long-term Constraintsmentioning

confidence: 99%

Section: B Oco With Long-term Constraintsmentioning

confidence: 99%

“…However, all of the above works on constrained OCO [10], [17]- [23] are under the standard per-time-slot update setting. No other known work considers periodic updates for OCO.…”

Section: B Oco With Long-term Constraintsmentioning

confidence: 99%