A Modular and Scalable Architecture for Reproducible Multi-Objective Optimization Experiments in Wireless Sensor Networks

Wireless Sensor Networks (WSNs) have diverse applications in urban, industrial and environmental monitoring. However, the design complexity of this type of network is high, due to conflicting objectives such as latency, energy consumption, connectivity and coverage. This article addresses the need for structured and reproducible approaches to developing WSNs. We propose a modular and scalable system designed to integrate simulators and evolutionary algorithms for multi-objective optimization in WSNs. We present a formalized process and supporting architecture that combines containerized simulations, a reactive data management layer, and a flexible optimization engine capable of handling diverse objective formulations and search strategies. The proposed environment enables distributed, simulation-based optimization experiments with automated orchestration, persistent metadata and versioned execution artifacts. To demonstrate feasibility, we present a prototype implementation that incorporates synthetic test modules and real WSN simulations using a classical simulator for simulating sensor networks. The results illustrate the potential of the proposed system to support reproducible and extensible research in design and optimization of WSNs.