Modelling Parallel Processing Methods for Resource Allocation under Nonlinear Dynamic Constraint

Doostmohammadian, Mohammadreza; Gabidulina, Zulfiya; Rabiee, Hamid R.

doi:10.22075/jme.2026.38634.2884

Modelling Parallel Processing Methods for Resource Allocation under Nonlinear Dynamic Constraint

Articles in Press

Document Type : Research Paper

Authors

mohammadreza doostmohammadian ¹

zulfiya Gabidulina ²

hamid R. Rabiee ³

¹ ُSemnan University

² Kazan Federal University, Russia

³ Faculty of Computer Engineering, Sharif University of Technology, Iran

10.22075/jme.2026.38634.2884

Abstract

This paper addresses the modeling of parallel processing methods in decentralized systems for resource allocation optimization. The growing interest in distributed and parallel algorithms stems from recent advances in the Internet of Things (IoT) and cloud computing. Distributed constrained optimization in multi-agent networks is proposed as an effective solution for large-scale applications, such as resource allocation optimization in smart power transmission networks and optimal scheduling of CPU processing networks in data centers. These decentralized algorithms offer advantages over centralized ones, including fault tolerance against single agent (node) failure, scalability, and improved performance in complex networks. Instead of analyzing data at a single processing node centrally, data is distributed among multiple distributed processing nodes and processed locally; subsequently, problem parameters are shared and coordinated in a collaborative distributed way. Also, we consider nonlinear constraints on the nodes/agents dynamic, such as quantization and saturation. The efficiency of this method has been validated through simulations in MATLAB.

Keywords

Decentralized Optimization

Resource Allocation

Parallel Processing

Complex Networks

Subjects