This study addresses the optimization of telecom network reliability in Ethiopia by applying numerical methods to improve network performance. Regularization and model selection criteria such as cross-validation are employed to optimise telecom networks in Ethiopia. Theoretical assumptions include the use of convex optimization problems, while key properties ensure stability and efficiency in network configurations. A specific proportion (60%) of simulated network failures were successfully mitigated by applying regularization techniques, demonstrating improved reliability. The findings suggest that regularization and cross-validation significantly enhance telecom network resilience in Ethiopia. Implementing these methods should be considered for future network upgrades to ensure robust performance. Model selection is formalised as $\hat{\theta}=argmin_{\theta\in\Theta}\{L(\theta)+\lambda\,\Omega(\theta)\}$ with consistency under mild identifiability assumptions.