In this paper, we propose a novel self-optimization framework that incorporates an automatic weight function for determining Handover Control Parameters (HCP) and utilizes the Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) algorithm for target cell selection. The proposed algorithm estimates suitable HCP values automatically based on key factors such as Signal-to-Interference-plus-Noise Ratio (SINR), traffic load, and user equipment (UE)'s speed. Additionally, TOPSIS algorithm is employed to select the most suitable target cell based on three evaluation criteria: SINR, Reference Signal Received Power (RSRP), and load level. The performance of the proposed algorithm is evaluated using load level, throughput, and Radio Link Failure (RLF) as key performance indicators (KPIs). Extensive simulations are carried out to assess the performance of the proposed approach. The findings indicate that the proposed algorithm outperforms the existing solutions in terms of load level, throughput, and RLF.

The caching paradigm has been introduced to alleviate backhaul traffic load and to reduce latencies due to massive never ending increase in data traffic. To fully exploit the benefits offered by caching, unmanned aerial vehicles (UAVs) and device-to-device (D2D) communication can be further utilized. In contrast to prior works, that strictly limits the content delivery routes up to two hops, we explore a multi-hop communications scenario, where the UAVs, the UEs, or both can relay the content to individual users. In this context, we formulate the problem for joint route selection and power allocation to minimize the overall system content delivery duration. First, motivated by the limitations of existing works, we consider the case where the nodes may transmit content simultaneously rather than sequentially and propose simple yet effective approach to allocate the transmission power. Second, we design a low-complexity greedy algorithm jointly handling route selection and power allocation. The simulation results demonstrate that the proposed greedy algorithm outperforms the benchmark algorithm by up to 56.98% in terms of content delivery duration while it achieves close-to-optimal performance.

Handover is the process of transferring a cellular call or data session from one base station (BS) to another. This process aims to establish reliable and uninterrupted connection, thereby providing satisfactory Quality of Service (QoS) and Quality of Experience (QoE) for users. 5G networks will use millimetre wave (mmWave) frequencies in addition to sub-6 GHz bands, which will make handover (HO) more challenging. This paper focuses on the problem of partially blind HOs which is a novel HO type. In this sense, we modify an extant algorithm used for the partially blind HOs [1] which the algorithm is based on machine learning (ML). In our modified algorithm, we use the extant algorithm with a powerful boosting method that is Categorical Boosting (CatBoost). We compare our modified algorithm with a baseline algorithm, the originally proposed algorithm, Support Vector Machine integrated original algorithm. Different settings of simulation time and number of users are considered in comparing the algorithms, and our modified algorithm outperforms the rest of the algorithms in majority of the settings with higher HO prediction rates as per obtained results. The obtained results clearly indicate that the integration of ML with partially blind HOs enables accurate predictions whether HO execution will be successful in collocated cells in a network in the real-world case. A noteworthy takeaway from the obtained results is that ML deployment with partially blind HOs will likely contribute to self-organizing networks (SONs) in 5G communication systems.