الفهرس 
Chapter 1 IntroductionOnly 14 pages are availabe for public view
 |  | from | 174 |  |  |
| | | from | 174 | | |
Abstract
As mobile networks continue to evolve to meet the rising demands of users, heterogeneous networks (HetNets) have emerged as a promising solution by integrating different technologies. The seamless handover of user connections within a HetNet, known as vertical handover (VHO), poses a challenge in achieving efficient performance from both network and user perspectives. This research proposes an adaptive optimized VHO approach, leveraging a multi-attribute decision-making (MADM) approach integrated with particle swarm optimization and gravitational search algorithm (PSOGSA). The proposed approach introduces a network selection framework that considers vital criteria such as signal strength, network attributes, user mobility, and service preferences. Additionally, two control parameters, Load Factor (LF) and Score Priority (SP), are introduced to minimize unnecessary handovers and reduce power consumption while ensuring balanced load distribution, then the desired objectives are formulated as an objective function and utilizes the PSOGSA algorithm to optimize LF and SP values during the handover process. Furthermore, we integrate machine learning (ML) techniques into the handover process by introducing a neural network (NN)-based VHO framework. The intelligent handover model exploits the capabilities of optimization to find out the optimal values of LF and SP parameters. First, a NN is trained using the data obtained from the optimization process based on changes in the network’s attributes and the number of involved users. Then, the trained NN is deployed to facilitate the handover decision-making process in HetNets, enabling the approach to adaptively learn from past experiences and make real-time informed decisions. This integration not only enhances optimization outcomes but also reduces computational complexity, expediting decision-making within HetNets and facilitating intelligent and efficient network management. To evaluate the proposed work, simulations are conducted using MATLAB R2019a in a HetNet environment, where fifth-generation (5G) wireless technology coexists with other radio access networks. The simulation results demonstrate that the proposed intelligent approach significantly reduces unnecessary handovers by over 60% and achieves improved load distribution of 20% to 40% compared to traditional handover approaches.