Towards an Original Unified Theory of Computational Resource Allocation for Heterogeneous Mobile Edge Computing in Nigerian Smart City Applications

Nigerian smart city applications, including traffic management, environmental monitoring, and public safety systems, increasingly depend on mobile edge computing infrastructure for latency-sensitive data processing, but the heterogeneous nature of Nigerian edge nodes, varying in computation capacity, power supply reliability, and network connectivity, demands a resource allocation theory that existing homogeneous edge computing models cannot provide. This study develops an original unified theory of computational resource allocation for heterogeneous mobile edge computing adapted to Nigerian smart city application requirements. The theory, designated H-MEC-NG, introduces three original theoretical constructs: a stochastic heterogeneous node availability model calibrated from empirical uptime and connectivity data collected from 24 edge nodes across three Nigerian cities over 90 days; a multi-objective resource allocation formalism that simultaneously minimises task latency, energy consumption, and task failure probability under heterogeneous reliability constraints; and a distributed online learning algorithm for adaptive resource scheduling that does not require centralised coordination, enabling the system to operate during partial network partitioning common in Nigerian urban environments. H-MEC-NG was validated through large-scale simulation using real workload traces from Abuja smart traffic management and Lagos environmental monitoring systems. Compared to homogeneous edge computing baselines, H-MEC-NG achieved 28 percent lower average task latency and 41 percent lower task failure rate under realistic Nigerian network conditions. The theory provides an original foundation for smart city edge computing design in resource-constrained urban environments.

Keywords: mobile edge computing, resource allocation, smart city, Nigeria, heterogeneous networks