동형암호 기반 군 의료 후송(MEDEVAC) 우선순위 결정 모델: 개인정보 보호 환경에서의 순위 보존 최적화 연구

초록

This paper proposes a privacy-preserving framework for prioritizing military medical evacuation (MEDEVAC) and ICU bed allocation in wartime and peacetime operations using Homomorphic Encryption (HE). Conventional MEDEVAC models rely on the plaintext sharing of clinical data, making it difficult to satisfy military security and medical privacy requirements. A polynomial logit structure tailored to HE constraints was designed, and the ranking stability of encrypted risk scores was theoretically analyzed. Sufficient conditions under which the priority set remains unchanged were derived by characterizing the relationship between the Top-boundary gap and encryption noise. The simulation results showed that, within specified noise bounds, the HE-based risk scores maintained a high rank correlation with plaintext scores and produced similar MEDEVAC and ICU allocation decisions, but excessive computational depth or accumulated noise may distort rankings. These findings highlight the feasibility of encrypted risk-based prioritization without direct access to individual clinical data, supporting a secure and operationally efficient decision framework for military medical command- and-control environments.

키워드