Comparative Analysis of User Association and Load Balancing Strategies in Cellular Networks: A MATLAB-Based Simulation Study

Abstract

Load imbalance is an important challenge in cellular networks because conventional user association mechanisms may produce uneven traffic distribution among neighboring base stations. This study compares three user association and load-balancing strategies using a MATLAB-based cellular network simulation. The investigated network consists of three base stations (BSs) and 300 users distributed over a 1000 × 1000 m² geographical area, with a maximum capacity of 100 Mbps assigned to each base station. Three scenarios are considered in the proposed comparative framework. The first scenario employs conventional nearest-base-station association and serves as the baseline. The second scenario applies a Decision Tree-based user association mechanism that incorporates network-related information into the association process. The third scenario employs a machine learning-based load balancing strategy designed to distribute users and traffic more evenly among the available base stations. The three strategies are evaluated under the same network configuration using base-station load, utilization level, user distribution, and load imbalance as comparative performance indicators. The simulation results illustrate the differences between distance-based association, load-aware Decision Tree association, and machine learning-based load balancing. Rather than assuming a predefined superior strategy, the study evaluates the behavior of each approach under identical simulation conditions and provides a MATLAB-based framework for comparative analysis of user association and load management strategies in cellular networks.