A MDP solution for Traditional Chinese medicine treatment planning

Herbal medicine is the primary method of treatment in Traditional Chinese medicine (TCM) which proposes an essential health solution in China. Medical treatments are usually made by TCM physicians sequentially in an uncertain environment. Markov Decision Process (MDP) provides a powerful mathematical technique for planning in environment under uncertainty and is suitable for TCM therapy planning. In this paper, we apply MDP to solve TCM herbal treatment planning with all the parameters inferred from TCM clinical data for patient with type 2 diabetes. This MDP model contains 30 health states obtained using k-means clustering algorithm and 159 actions of basic prescriptions. This model could order sequences of prescriptions from the action set for patients with type 2 diabetes. The results show that the MDP model for TCM treatment planning can identify and order useful prescriptions which are reasonable in clinical practice.