Constrained Multi-Objective Routing Optimization in Distributed Networks Using Classical Iterative Methods

Core Focus: This paper investigates the systematic application of four classical iterative optimization algorithms (Steepest Descent, Newton’s Method, Quasi-Newton BFGS, and Conjugate Gradient) to solve constrained multi-objective routing problems in distributed networks. Problem & Constraints: It models routing optimization using four decision variables (route selection, bandwidth allocation, load distribution, and queue priority) to minimize latency, cost, and packet loss while maximizing network reliability under strict capacity, latency, load balancing, and QoS constraints. Key Findings: Tested across four network scenarios (normal operation, peak traffic, network failures, and heterogeneous traffic), the study demonstrates that second-order methods drastically outperform first-order techniques, with Quasi-Newton BFGS emerging as the most practical and robust overall choice due to its high convergence rate and low computational overhead