A geometry-driven fixed small-cell deployment framework with adaptive user association for mmWave networks

Efficient small-cell deployment and balanced traffic distribution are essential for improving the performance of heterogeneous millimeter-wave (mmWave) networks. This paper proposes a Fixed Geometric Shape Deployment Arrangement (FGSDA), a deterministic geometry-based deployment framework that positions pico base stations using regular polygon configurations to achieve uniform spatial distribution, predictable deployment geometry, and repeatable network planning. Unlike adaptive deployment approaches that iteratively optimize base-station locations, the proposed framework maintains fixed small-cell positions while adapting traffic distribution through a Lagrangian-based user association algorithm. A geometric feasibility criterion is incorporated to ensure that all deployed pico base stations remain within the macrocell coverage region before network evaluation. The proposed framework is evaluated for two representative deployment scenarios, namely UAV-assisted mobile small cells and indoor shopping mall networks, using standardized 3GPP mmWave propagation models operating at carrier frequencies of 28, 38, 60, and 73 GHz. Performance is assessed through 100 independent Monte Carlo simulations using the Load Balancing Index (LBI), average Signal-to-Interference-plus-Noise Ratio (SINR), average network throughput, spectral efficiency, coverage probability, outage probability, average interference, energy efficiency, average end-to-end delay, and rejected traffic ratio. The proposed FGSDA is compared with representative deterministic deployment strategies, including Grid, Hexagonal, Random, and Quadrant-Based deployments, under identical simulation conditions. The simulation results demonstrate that the proposed deterministic deployment consistently provides superior load balancing and communication performance across both deployment scenarios, achieving higher LBI, SINR, throughput, spectral efficiency, coverage probability, and energy efficiency while reducing outage probability, average interference, end-to-end delay, and rejected traffic ratio compared with all evaluated baseline deployment methods. Within the geometrically feasible deployment configurations considered in this study, the eight-pico-base-station configuration provides the highest LBI, while the indoor shopping mall scenario operating at 60 GHz achieves the best overall communication performance among the evaluated carrier frequencies. These results demonstrate that deterministic geometric deployment combined with adaptive user association provides an effective and computationally efficient framework for planned heterogeneous mmWave networks while maintaining deployment repeatability and practical applicability.