In this article, we investigate the load balancing in the era of network slicing in radio access network (RAN). In our differentiated coverage RAN slicing scenario, different coverage areas of different slices have an impact on user association with base station (BS) and users’ resources obtained from BSs. Therefore, slices will limit the balance of BSs’ load. Based on this, we mainly construct a performance index to measure load balancing of RAN slicing, namely load balancing utility function. We formulate a joint optimization problem of user association and radio resource allocation for maximizing utility function. To solve this problem, we simplify and decompose it into three sub-problems, user association, slice resource allocation, and user resource allocation. We propose an iterative algorithm that uses Lagrange duality-based, Lagrange multiplier method-based, and uniform distribution methods to solve the sub-problems respectively. Finally, we obtain an approximate optimal solution with low complexity. The simulation results show that our proposed algorithm can effectively improve the load balancing utility compared with the benchmark algorithms, which means it gets the most reasonable user association and the best resource allocation.

在本文中，我们研究了无线接入网 (RAN) 中网络切片时代的负载平衡。在我们的差异化覆盖RAN切片场​​景中，不同切片的不同覆盖区域会影响用户与基站（BS）的关联以及从BS获得的用户资源。因此，切片将限制BS负载的平衡。在此基础上，我们主要构建了一个衡量RAN切片负载均衡的性能指标，即负载均衡效用函数。我们为最大化效用函数制定了用户关联和无线电资源分配的联合优化问题。为了解决这个问题，我们将其简化分解为三个子问题，用户关联、切片资源分配和用户资源分配。我们提出了一种使用基于拉格朗日对偶的迭代算法，基于拉格朗日乘子法和均匀分布法分别求解子问题。最后，我们得到一个低复杂度的近似最优解。仿真结果表明，与基准算法相比，我们提出的算法可以有效提高负载均衡效用，这意味着它获得了最合理的用户关联和最佳的资源分配。