Enhancing Data Collection in Heterogenous Wireless Sensor Networks: A Novel Tree-Structured Genetic Algorithm Approach

Heterogenous Wireless Sensor Networks (HWSNs) are cost-effective solutions to monitor the environment. These networks face limited battery as a critical challenge. Traditionally, energy-efficient routing algorithms have been the primary solution. While Artificial Intelligence (AI)-based approaches, particularly metaheuristics, show promise, current methods suffer from a key limitation: problem formulation hinders their success. Existing approaches rely on the array data structure for tree construction, leading to inefficiency. This paper proposes a novel Genetic Algorithm (GA)-based approach for data collection in HWSNs. It introduces a simpler and more intuitive tree data structure specifically designed by GA, leading to improved performance. Additionally, the paper proposes population initialization schemes and GA operators customized for the problem. Finally, an advanced cost function is employed that distributes workloads based on hop count to the Base Station (BS), further optimizing energy consumption. Extensive simulations demonstrate the superiority of the proposed algorithm in data gathering. It increases the number of received bits to the BS by an average of 6.6%, while also almost significantly improves the network lifetime and the number of alive nodes.