This project falls under the interdisciplinary research and basic application research of traditional Chinese medicine (TCM) and pharmaceutical informatics. Homemade preparations of TCM are a characteristic and advantage of traditional Chinese medical hospitals, serving as the cradle for the development of new TCM drugs. However, the general status of hospital preparations is characterized by low scientific and technological content, outdated processes, and relatively low quality control levels. By applying new theories and methods from interdisciplinary fields, research on creating complex TCM group-effect relationship models and computer-aided optimization design technology for new TCM drugs will help to solve key issues related to TCM preparation research. This project applies pharmaceutical informatics technology to optimize existing formulations, improve preparation processes, control the quality of TCM, and search for new candidate prescription agreements throughout the entire process of developing hospital TCM preparations, providing new ideas and methods for the development and research of hospital preparations. In terms of current research on hospital preparations, Using a new research strategy of 'hypothesis formulation → experimental design → information integration → computational modeling → identification, prediction, optimization → experimental verification', this project has conducted formula optimization, process optimization, and quality control of hospital preparations. It provides beneficial ideas for effectively addressing some key issues faced by traditional Chinese medicine (TCM) preparations at the current stage: In terms of formula optimization, this project proposed a new strategy involving uniform design, pharmacodynamic testing, multi-index comprehensive evaluation, LASSO algorithm modeling, evolutionary search (genetic algorithm) optimization, and experimental verification for the lipid-lowering granules of our hospital. In terms of optimizing the preparation process, this project proposed a multi-index optimization solution using a BP neural network combined with a genetic algorithm. This approach overcomes the disadvantage of needing to set weights in advance when optimizing for a single objective and has been successfully applied to the process optimization of Qinbing Eye Drops. An AHP-CRITIC multi-index weight solution was proposed and successfully applied to the process optimization of Compound Ziyuqing Granules. In terms of controlling the quality of traditional Chinese medicine, for the evaluation of raw medicinal materials such as Astragalus membranaceus and Cortex Fraxini, this project proposed the application of comprehensive principal component analysis and AHP-CRITIC to the quality comprehensive evaluation of these raw materials, finding that both methods could comprehensively represent the information of the original samples, and the results of the quality ranking were relatively objective. A new strategy for optimizing the quality of formulation extracts was proposed, which combined fingerprint similarity with indicator component content to optimize the quality of Cortex Fraxini extractions. This project developed some methods using ultraviolet spectroscopy in combination with intelligent computational models, such as artificial neural networks (ANN) and genetic algorithms combined with least squares support vector machines (GA-LSSVM). After achieving rapid analysis of the quality of preparations such as Benzydamine Nasal Drops and Qinbing Eye Drops, we have made significant progress in the discovery of new hospital preparation candidate prescriptions. Since hospital preparations mainly come from hospital agreement prescriptions or empirical formulas of famous traditional Chinese medicine practitioners, they are accumulated and summarized by our hospital's TCM doctors based on the principles of syndrome differentiation and treatment over a long period of time. To ensure the sustainable development of hospital preparations, new ones must be derived from new effective agreement prescriptions. Guided by pharmaceutical informatics technology, this project has proposed indicators for assessing the core nature and effectiveness of prescriptions, as well as methods for discovering core effective prescriptions in the medical practice of traditional Chinese medicine. These methods include a discovery process based on complex networks, from core drugs (combined with centrality analysis) → core drug pairs (combined with information entropy) → core prescriptions (combined with BK algorithm) → core effective prescriptions (combined with GA algorithm). The project has successfully summarized the medication experience formulas of renowned experts in traditional Chinese medicine rheumatism, dermatology, oncology, and lung diseases from our hospital, including Professor Chen Xiangjun, Professor Ma Shaoyao, Professor Liu Jiashang, and Professor Shao Changrong. In the medical practice of the Oncology Department, it has identified several potentially effective core prescriptions that can serve as candidate formulas for the development of new in-house preparations. The project has published 27 academic papers, including 4 SCI articles, with a total of 144 citations, of which 130 are from other sources, and one SCI article was cited 5 times; it has also trained 6 graduate students. This has formed the characteristic of interdisciplinary research on traditional Chinese medicine formulations in the hospital and has cultivated a research team that is brave in innovation.