<p>To address the growing demand of indoor wireless communication, hybrid networks that integrate visible light communication (VLC) and radio frequency (RF) communication have emerged as a promising solution. However, in VLC-RF hybrid networks, mismatches between supply and demand make it difficult to meet user quality of experience (QoE), often resulting in either severe QoE degradation or resource over-provisioning. To overcome this challenge, relying on the framework of semi-grant-free (SGF) random access, a joint optimization of resource allocation (JORA) scheme based on the improved Rüppell’s fox optimizer (IRFO) is proposed, which aims to satisfy different user QoE requirements with minimal resource consumption. First, a VLC-RF hybrid network model is established, and users are grouped according to VLC coverage and their QoE requirements. The RF compensation mechanism is then designed to mitigate potential VLC service deficiency, and the effective throughput is calculated for each user. Furthermore, a QoE evaluation model is formulated in view of the mean opinion score (MOS) mapping function, and the QoE requirements are quantified for different users. With the objective of minimizing RF service rate, an optimization problem with multiple constraints is set up. Finally, the Rüppell’s fox optimizer (RFO) is enhanced with logistic mapping (LM) and differential evolution (DE), and the IRFO-JORA algorithm is developed to optimize network resource allocation. Simulation results demonstrate that the IRFO-JORA algorithm could efficiently allocate resource according to system parameters and user QoE requirements, thereby significantly improving RF utilization efficiency.</p>