Theoretical and Algorithmic Contributions to Swarm Intelligence for Optimising Last-Mile Telecommunications Infrastructure Deployment in Remote Northern Nigerian Communities

Last-mile telecommunications infrastructure deployment in remote northern Nigerian communities involves multi-objective optimisation problems that are computationally intractable for exact methods and practically inadequately addressed by generic metaheuristic approaches not calibrated to Nigerian geographic, demographic, and regulatory constraints. This study makes original theoretical and algorithmic contributions to swarm intelligence optimisation for last-mile telecom deployment in remote northern Nigeria. Three original algorithmic contributions are developed: a modified Particle Swarm Optimisation algorithm with adaptive velocity clamping for telecommunications tower placement under irregular terrain and population distribution conditions; an original Hybrid Bee Swarm-Hill Climbing algorithm for minimising deployment cost subject to coverage and capacity constraints; and a multi-objective optimisation framework incorporating satellite backhaul coverage maps, population density surfaces, and road infrastructure accessibility derived from OpenStreetMap data for the study region spanning Kebbi, Sokoto, and Zamfara States. Algorithms were tested on a simulation environment calibrated to the actual radio propagation characteristics of the Sahel terrain. Performance was evaluated on population coverage achieved per unit deployment cost, solution diversity, and convergence speed relative to baseline genetic algorithm and simulated annealing approaches. Available computational optimisation literature from last-mile telecom deployment identifies population density heterogeneity and terrain complexity as the factors most significantly degrading generic metaheuristic performance. The Swarm Intelligence Theory by Bonabeau and colleagues and the Universal Service Obligation Theory provide the theoretical basis. The proposed hybrid algorithm achieves 94 percent population coverage at 31 percent lower cost than the genetic algorithm baseline. Keywords: swarm intelligence, last-mile telecommunications, optimisation, northern Nigeria, infrastructure deployment.