Intelligent wireless network selection using linguistic fuzzy neural networks

Wireless networks are fundamental to modern digital communication, and selecting the most appropriate network among competing technologies is a complex multi-criteria group decision-making problem. The evaluation process involves several conflicting attributes, such as bandwidth, transmission delay, cost, and data loss rate, and becomes more challenging when expert assessments are expressed in uncertain linguistic terms. To address these issues, this study develops a novel linguistic pq-rung orthopair fuzzy neural network (Lpq-ROFNN) model integrated with linguistic pq-rung orthopair Dombi fuzzy weighted aggregation operators. The proposed framework enables decision-makers to express evaluations using linguistic pq-rung orthopair fuzzy sets, offering greater flexibility and improved uncertainty management compared to traditional fuzzy models. The model is applied to a heterogeneous wireless network selection problem involving eight alternatives evaluated over eight attributes by four experts. Entropy measures are employed to determine unknown attribute weights objectively, while Dombi-based aggregation operators combine expert information to compute hidden layer outputs within the neural network structure. Final scores are obtained through a score function and sigmoid activation process. The results identify $${p}_{1}=$$ 5G NR as the most suitable network due to its superior bandwidth, low latency, reliability, and efficient performance. Sensitivity analysis with respect to Dombi and $$(p, q)$$ parameters confirms ranking stability. Comparative analysis, including extended TOPSIS and other MCGDM methods, demonstrates high consistency and robustness, highlighting the model’s effectiveness, scalability, and practical applicability in complex decision-making environments.