Fuzzy-AHP 분석을 이용한 국내 AI 데이터센터 입지요인의 중요도 분석

초록

As AI services and generative AI applications rapidly expand, the importance of selecting suitable locations for AI data centers has increased to support large-scale computing facilities and stable AI service provision. This study aims to identify evaluation factors for AI data center location decisions in Korea and empirically analyze their importance and priorities by applying AHP and Fuzzy-AHP based on Chang’s extent analysis method. A hierarchical structure was established with four main criteria, nine sub-criteria, and 33 detailed factors, based on prior studies and expert review. A survey was conducted with experts in related fields, and 28 valid responses were used for the analysis. The results indicate that infrastructure and resource connectivity ranked highest among the main criteria in both AHP and Fuzzy-AHP, followed by network locational foundation, governance and social operational conditions, and environmental risk sustainability. At the detailed-factor level, grid capacity, timely grid connection, water supply stability, power supply stability, and power backup capability ranked highly in the AHP results, while grid capacity, timely grid connection, power supply stability, power backup capability, and water supply stability ranked highly in the Fuzzy-AHP results. These findings indicate that power grid conditions and cooling environment conditions should be prioritized in AI data center location decisions in Korea. This study is significant in that it reflects domestic locational conditions and empirically analyzes the priorities of AI data center location factors in Korea, where related empirical research remains limited.

키워드