Performance evaluation and optimization for 6G networks: A survey of KPIs, Tools, and AI models

This paper presents a comprehensive review of performance evaluation and optimization (PEO) methodologies for sixth-generation (6G) networks. It systematically analyzes the key performance indicators (KPIs) defined by major standardization bodies such as ITU-R, 3GPP, IEEE, & TIA and industry fora such as 5G-PPP, 6G-RIC, NGMN, and 6G-IA. Beyond summarizing KPI targets, this study classifies evaluation tools and AI-driven optimization frameworks into a unified taxonomy, highlighting how artificial intelligence enables adaptive, sustainable, and cross-layer performance control. Comparative synthesis reveals that most AI-driven approaches remain simulation-centric, lack benchmark alignment across KPIs, and face challenges in scalability, trustworthiness, and real-time inference. The paper concludes by identifying research gaps in AI-driven PEO frameworks and emphasizing future directions toward autonomous, energy-efficient, and explainable 6G systems.