이태식 연구실

Research on Partially Controllable Multi-Agent Systems (PCMAS) where only a subset of agents can be directly controlled. Develops encoder-decoder-based deep reinforcement learning framework with self-attention mechanism to capture inter-agent interactions and design coordinated policies.

Elucidates patient hospital choice behaviors using Discrete Choice Model (DCM). Introduces Latent Multi-class Choice Model (LMCM) to capture heterogeneity in utility functions and uses Nested Logit model to derive hierarchical hospital structure from patient perspective.

Develops game theory-based mathematical model to represent patient concentration at large hospitals. Derives Nash equilibria to explore effects of various policies including penalties, direct investment, and pay-for-performance systems.

Redefines healthcare service areas for secondary and tertiary care in Korea using Relevance Index (RI) and Commitment Index (CI). Proposes revised framework of 60 healthcare service areas based on medical utilization patterns and demographic changes.

Proposes analytical framework extracting sequential medical usage events from NHIS claims data to construct patient sharing network. Applies community detection algorithms to identify hospital clusters and analyze relationships based on function and location.

Addresses healthcare supply structure imbalance in Gwangju and Jeonnam regions. Employs mathematical optimization model to identify optimal enhancement of existing hospitals' functions to better meet demand and improve patient outcomes.