Akademik Veri Yönetim Sistemi
Pervasive and Mobile Computing, cilt.120, 2026 (SCI-Expanded, Scopus)
This study presents a sustainability-driven integrated multi-objective optimization (MOO) approach for configuring wireless sensor networks (WSNs), with the dual objective of extending network lifetime and minimizing end-to-end delay under selective α-coverage and connectivity constraints. By defining the most efficient network state, the model provides a rigorous methodology for mitigating energy dissipation, thereby advancing the core principles of green communications and sustainable infrastructure. The selective α-coverage strategy is integrated into the model to ensure strategically distributed sensing coverage, thereby reducing redundant energy expenditure through intelligent, rather than exhaustive, coverage. The resulting problem is expressed as a non-linear 0–1 mixed-integer programming (MIP) model and is reformulated into two linear equivalents for improved computational tractability. In addition, a third highly efficient linear model is developed by transforming the lifetime maximization objective into a maximum energy consumption minimization objective. An extensive computational study is conducted on randomly generated grid-based WSN instances with three grid sizes and three deployment densities, where 30 instances are solved for each configuration using Gurobi. Network lifetime, average delay and total packets received are compared across alternative MILP formulations, and the performance of the models is examined through distributional analyses and non-parametric significance tests. Experimental results demonstrate that the proposed bi-objective optimization approach more than doubles the lifetime of the network compared to delay-centric designs and achieves a 35% reduction in the average delay relative to lifetime-oriented models. Furthermore, we incorporate explicit control over the number of sensors deployed into the optimization model to address improved resource efficiency and better capture environmental considerations. The test results show that it is possible to maintain optimal network lifetime and average delay levels by using up to 50% fewer sensors.