This paper addresses the problem of user association in heterogeneous cellular networks (HetNets) to achieve load balancing. The goal is to improve network performance by actively pushing mobile users to less loaded small cells (e.g., pico and femtocells), even if they offer a lower instantaneous SINR than the macrocell. The paper proposes a distributed algorithm that converges to a near-optimal solution with a theoretical performance guarantee. It also shows that simple per-tier biasing can achieve significant performance gains, with a 3.5x throughput gain for cell-edge users and a 2x rate gain for median users compared to max-SINR association. The paper investigates optimal and near-optimal solutions for user association in HetNets, particularly those with simple coordination requirements. It considers both load-aware and non-load-aware association schemes. The paper introduces a logarithmic utility function, which provides a desirable trade-off between opportunism and fair allocation across users. It also proposes a distributed algorithm based on dual decomposition, which allows for efficient and low-overhead implementation in HetNets. The paper also evaluates the performance of different association schemes, including max-SINR, fractional association, and biasing. It shows that the proposed load-aware association scheme significantly improves resource utilization and mitigates macrocell congestion, resulting in a multi-fold gain in overall rate for most users, especially those with previously low rates. The paper also evaluates the effects of BS density and transmit power on biasing factors, showing that optimal biasing factors are nearly independent of BS densities but highly dependent on per-tier transmit powers. The results demonstrate that the proposed load-aware association scheme achieves near-optimal performance with low complexity.This paper addresses the problem of user association in heterogeneous cellular networks (HetNets) to achieve load balancing. The goal is to improve network performance by actively pushing mobile users to less loaded small cells (e.g., pico and femtocells), even if they offer a lower instantaneous SINR than the macrocell. The paper proposes a distributed algorithm that converges to a near-optimal solution with a theoretical performance guarantee. It also shows that simple per-tier biasing can achieve significant performance gains, with a 3.5x throughput gain for cell-edge users and a 2x rate gain for median users compared to max-SINR association. The paper investigates optimal and near-optimal solutions for user association in HetNets, particularly those with simple coordination requirements. It considers both load-aware and non-load-aware association schemes. The paper introduces a logarithmic utility function, which provides a desirable trade-off between opportunism and fair allocation across users. It also proposes a distributed algorithm based on dual decomposition, which allows for efficient and low-overhead implementation in HetNets. The paper also evaluates the performance of different association schemes, including max-SINR, fractional association, and biasing. It shows that the proposed load-aware association scheme significantly improves resource utilization and mitigates macrocell congestion, resulting in a multi-fold gain in overall rate for most users, especially those with previously low rates. The paper also evaluates the effects of BS density and transmit power on biasing factors, showing that optimal biasing factors are nearly independent of BS densities but highly dependent on per-tier transmit powers. The results demonstrate that the proposed load-aware association scheme achieves near-optimal performance with low complexity.