Application of the adaptive sparrow search algorithm in medical supply engineering
Yuchi Li, Zihao Song, Lili Xia, Jiqing Sun, Ziqing Wang
Abstract
To enhance the cost-effectiveness of vascular robotic systems in clinical settings, this study constructs an integrated forecasting-optimization framework for long-term resource planning. A weekly demand forecasting model is developed using the SARIMA approach, with model order selection guided by stationarity testing and ACF and PACF analysis. Forecast accuracy is validated to ensure reliable downstream optimization. Based on the predicted 112-week demand, a nonlinear procurement scheduling model is formulated, incorporating Poisson-distributed scrap rates and two types of acquisition strategies: conventional and emergency-use procurement. To solve the mixed-integer nonlinear programming problem under constraints such as maintenance cycles, training limits, and resource coupling, a continuous relaxation method is adopted along with a penalty-based cost function. The problem is then optimized using an Adaptive Sparrow Search Algorithm (ASSA), enhanced with Levy flights and adaptive producer ratios. Extensive sensitivity and interaction analyses are conducted on parameters including scrap rate, training limits, and initial inventory levels. The results not only demonstrate the robustness of the proposed approach but also offer valuable insights into strategic procurement under dynamic clinical demand, providing a novel data-driven paradigm for hospital resource allocation.