Heterogeneous cellular networks with hybrid energy supplies can relieve traffic pressure and reduce grid energy consumption. In heterogeneous cellular networks, rational resource management can help improve system performances. In general, more than one performance is expected to do well, but there can exist a trade-off among different performance metrics, thus making resource management a multi-objective problem. The existing solution usually transforms a multi-objective problem into another single-objective problem by assigning weights for various objectives. However, it is difficult to know the exact weights in advance, and different systems call for different requirements for objectives. Hence, a multi-objective optimization approach based on the gravitational search algorithm (GSA) is proposed to find a series of Pareto optimal solutions. The decision-makers can select an appropriate solution according to the system requirement. In this work, three different multi-objective GSA-based algorithms are proposed to determine user association and power control, with the goal to optimize the traffic load balancing among small base stations and grid energy consumption per unit throughput simultaneously. The complexity of the proposed algorithms is analyzed, and simulations compare the performances of the proposed algorithms and the benchmark algorithm. Experimental results reveal the feasibility and effectiveness of this approach.

中文翻译：

---

具有混合能源供应的异构蜂窝网络的资源管理：一种多目标优化方法

具有混合能源供应的异构蜂窝网络可以缓解交通压力并降低电网能耗。在异构蜂窝网络中，合理的资源管理有助于提高系统性能。一般来说，预期不止一个性能会做得好，但不同性能指标之间可能存在权衡，从而使资源管理成为一个多目标问题。现有的解决方案通常通过为各种目标分配权重将多目标问题转化为另一个单目标问题。但是，很难提前知道确切的权重，不同的系统对目标的要求也不同。因此，提出了一种基于引力搜索算法（GSA）的多目标优化方法来寻找一系列帕累托最优解。决策者可以根据系统需求选择合适的解决方案。在这项工作中，提出了三种不同的基于多目标 GSA 的算法来确定用户关联和功率控制，目的是同时优化小基站之间的流量负载平衡和单位吞吐量的电网能耗。分析了所提算法的复杂度，并通过仿真比较了所提算法与基准算法的性能。实验结果表明了这种方法的可行性和有效性。目标是同时优化小基站之间的流量负载平衡和单位吞吐量的电网能耗。分析了所提算法的复杂度，并通过仿真比较了所提算法与基准算法的性能。实验结果表明了这种方法的可行性和有效性。目标是同时优化小基站之间的流量负载平衡和单位吞吐量的电网能耗。分析了所提算法的复杂度，并通过仿真比较了所提算法与基准算法的性能。实验结果表明了这种方法的可行性和有效性。