Optimizing Resource Allocation in 5G Network Slicing: A Hybrid MLR- LP Approach to Minimize Latency in Ride-Hailing Services

Abstract

The advent of 5G networks has ushered in a new era of
communication technology characterized by unprecedented speed,
ultra-low latency, and higher reliability. However, intelligent and
adaptive resource allocation in 5G network slicing is critical to
meeting consistent sub-10 ms latency, a value that aligns with the
performance
benchmark
of
ultra-reliable
low-latency
communications such as ride-hailing. This study deployed a hybrid
Multiple Linear Regression–Linear Programming (MLR-LP)
framework for optimizing bandwidth, memory, and signal strength
to achieve latency reduction. Real-time data were collected from
ride-hailing sessions in 5G-covered areas of Benin City, Nigeria,
capturing latency, bandwidth, memory, and signal strength. The
MLR model established the predictive relationship between
network resources and latency, achieving a strong R² value of
0.941. The regression equation was embedded as the objective
function of an iterative LP model, which optimized bandwidth,
memory, and signal strength allocations. The iteration process was
guided by practical feasibility and variability analysis, particularly
the unit step standard deviation, to progressively expand the bounds
of resource variables in a controlled manner until feasible sub-10
ms latency was consistently obtained. The results demonstrate that
the variability-driven iterative MLR-LP approach effectively
minimizes latency to reliably support latency-sensitive services and
enhancing 5G slicing performance. The study concludes that
integrating predictive modeling with optimization techniques
provides both theoretical and practical contributions, offering a
possible solution for adaptive 5G resource management.