Downlink (DL)-uplink (UL) coupled cell association scheme was adopted in 3rd generation (3G) homogeneous network and 4th generation (4G) heterogeneous network (HetNet). In the coupled cell association scheme, a user is associated to single base station (BS) in DL & UL based on the strongest received signal power (SRSP) in DL from a macro base station (MBS) and multiple small base stations (SBS) in HetNet. This is a sub-optimal solution for cell association as most of the users are associated to a MBS due to dominant transmit power and brings challenges like multiple interference issues and imbalanced user traffic load which leads to a degraded throughput in HetNet. In this paper, we investigate downlink uplink decoupled cell association scheme to address these issues and formulate a sum-rate maximization problem in terms of admission control, cell association and power allocation for MBS only, coupled and decoupled HetNet. The formulated optimization problem falls into class of a mixed integer non linear programming (MINLP) problem which is NP-hard and requires exhaustive search to find the optimal solution. However, computational complexity of the exhaustive search increases exponentially with the increase in number of users. Therefore, an outer approximation algorithm (OAA) is proposed as a solution to find near optimal solution with less computation complexity. Extensive simulations work has been done to evaluate the decoupled cell association scheme in HetNet vs the coupled cell association scheme in traditional MBS only and HetNet. Results show the effectiveness of decoupled cell association scheme in terms of KPIs, such as throughput, addressing user traffic load imbalances and number of users attached.

第三代（3G）同质网络和第四代（4G）异构网络（HetNet）采用下行（DL）-上行（UL）耦合小区关联方案。在耦合小区关联方案中，基于DL中来自宏基站（MBS）和多个小型基站（SBS）的最强接收信号功率（SRSP），将用户与DL和UL中的单个基站（BS）关联在HetNet中。这是用于小区关联的次优解决方案，因为大多数用户由于主要的发射功率而与MBS关联，并带来诸如多重干扰问题和不平衡的用户业务负载之类的挑战，从而导致HetNet吞吐量下降。在本文中，我们研究了下行链路上行链路解耦小区关联方案，以解决这些问题，并在准入控制方面制定总和最大化问题，仅针对MBS，耦合和解耦的HetNet的小区关联和功率分配。拟定的优化问题属于NP整数的混合整数非线性规划（MINLP）问题，需要详尽搜索以找到最佳解。但是，穷举搜索的计算复杂度随着用户数量的增加而呈指数增长。因此，提出了一种外部近似算法（OAA）作为一种解决方案，以较少的计算复杂度找到接近最优的解决方案。已经进行了广泛的仿真工作，以评估HetNet中的解耦单元关联方案与仅传统MBS和HetNet中的耦合单元关联方案。结果显示，根据KPI（例如吞吐量，