.png){kind=icon}
.png){kind=icon}
Return
Enabling Multi-Application Multi-Objective Autoscaling With Quick Analytic Hierarchy Process
.png){kind=icon}
PRE
AI
DOI：10.1109/LNET.2026.3657848.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Dynamic autoscaling in Edge Clusters must reconcile conflicting resource demands from multiple co-hosted applications within short decision intervals, as conventional single-application controllers fail to consider such inter-application competition. The computational cost of classic optimization and decision-making methods increases rapidly with the number of alternative resource deployments, limiting their real-time applicability for short control loops typical in 5G/6G environments with rapid demand fluctuations. This letter introduces quick Analytic Hierarchy Process (qAHP), a modification of AHP formulation that preserves the ranking consistency of AHP while reducing the priority vector computation to linear time with respect to the number of alternatives. We formally prove this complexity reduction, achieving identical decision outcomes to classic AHP, and we validate it through proof-of-concept experiments showing that qAHP enables real-time autoscaling, mitigates over- and under-provisioning, and achieving up to two orders of magnitude lower execution time compared to classic AHP.
Keywords:
Vectors
Complexity theory
Phase change materials
Real-time systems
Optimization
Monitoring
Dynamic scheduling
Resource management
MCDM
Autoscaling
resource management
complexity reduction
analytic hierarchy process
analytic hierarchy process
Journal
I
IF:
0
Papers:
62
Citations:
0