Document Type : Original/Review Paper
Authors
Department of Computer Engineering, Shahrood University of Technology, Shahrood, Iran
Abstract
In the era of rapid internet expansion, Edge Computing has emerged as a critical field in computer science, addressing challenges posed by increasing internet data, bandwidth limitations, and the distance between cloud servers and users. This paper presents a Multiple Knapsack Problem (MKP)-based two-tier approach for optimizing resource management in hybrid edge and cloud computing environments. The proposed methodology models servers as knapsacks with limited capacity and tasks as items with weight and value, employing a weighted dimensionality reduction technique that transforms multi-dimensional server attributes into a single-dimension capacity measure and multi-dimensional task attributes into a two-dimension value/weight representation. The MKP is solved using an exact Branch-and-Bound algorithm for small-scale instances and a greedy heuristic for larger instances. The approach is evaluated using the ECHOES simulation environment with 20 cloud and edge servers and 100 task sets. Results demonstrate that the MKP-based strategy achieved the highest mean successful task count across all evaluated offloading strategies, including Decision Tree, EdgeFirst, Random, and Default baselines, with statistical significance (p<0.001). The findings confirm that the knapsack-based formulation provides an effective and computationally tractable framework for task offloading decisions in edge computing systems.
Keywords
Main Subjects
)