Stabilizing a Queue Subject to Activity-Dependent Server Performance

We consider a discrete-time system, comprising a first-come-first-served queue, a nonpreemptive server, and a scheduler that governs the assignment of tasks in the queue to the server. The server has an availability state that indicates, at each instant, whether the server is busy working on a task...

Full description

Saved in:
Bibliographic Details
Published in:IEEE transactions on control of network systems Vol. 8; no. 4; pp. 1579 - 1591
Main Authors: Lin, Michael, La, Richard J., Martins, Nuno C.
Format: Journal Article
Language:English
Published: Piscataway IEEE 01.12.2021
The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
Subjects:
Availability
Discrete time systems
Policies
Queueing analysis
Queues
Scheduling algorithms
Servers
Stability
Task scheduling
ISSN:2325-5870, 2372-2533
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:We consider a discrete-time system, comprising a first-come-first-served queue, a nonpreemptive server, and a scheduler that governs the assignment of tasks in the queue to the server. The server has an availability state that indicates, at each instant, whether the server is busy working on a task or is available. In the latter case, if the queue is nonempty, then a task-assignment control policy implemented by the scheduler either assigns a new task to the server or allows it to rest. The server also has an integer-valued activity state that is nonincreasing during rest periods, and is nondecreasing otherwise. An instantaneous service rate function ascribes to each possible value of the activity state a probability that the server can complete a task in one time step. For a typical instantaneous service rate function, the completion probability decreases (server performance worsens) as the activity state increases. The scheduler policy has access to the queue size and the entire state of the server. In this article, we study the problem of designing scheduler policies that stabilize the queue. We show that stability, whenever viable, can be achieved by a simple policy that bases its decisions on the availability state, a threshold applied to the activity state, and a flag that indicates when the queue is empty. The supremum of the service rates achievable by stabilizing policies can be determined by a finite search. Our results remain valid even when the instantaneous service rate function is not monotonic.
Bibliography:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:2325-5870
2372-2533
DOI:10.1109/TCNS.2021.3074527