Objective: The aim of this study was to optimize patient flow to reduce carbon dioxide emissions by considering the waiting time, length of stay and operational capacity of the treatment system.
Methodology: In this study, the methodology includes three methods for optimizing carbon dioxide in the patient flow. The first method is simulation-optimization. The second method is a combination of data mining and simulation-optimization. The third method is the development of the Chaotic Salp Swarm evolution algorithm with opposition-based learning algorithm (OBL) and sine cosine algorithm (SCA). This extension is called the CSSAOS evolution algorithm. In the first method, a discrete event simulation model was generated. In addition, OptQuest optimization method was used to optimize the simulation model. In the second method, four clustering algorithms (K-means, K-medoid, hierarchical clustering, and fuzzy C-means) for clustering patients based on age, sex, treatment cost, length of stay (LOS), CABG, and pPCI / PCI was used. In the third method, a patient flow optimization model was developed along with minimizing the amount of carbon dioxide emitted in the patient flow. The model is then solved using the proposed CSSAOS algorithm.
Results: Simulation-optimization outputs of one-objective problem solving lead to a decrease of 0.40% in carbon dioxide emissions, 56.6% in waiting time, 0.42% in residence time, and 2.18% increase in patient throughput. In addition, solving the multi-objective problem using ε-constraint and OptQuest shows that it is not possible to keep the amount of carbon dioxide produced at the minimum single-objective level if it is important for the hospital to achieve other goals. The minimum amount of carbon dioxide produced is 130,139 kg. If the hospital seeks to reduce the length of stay and waiting time for patients, then the result will be an increase in carbon dioxide produced to 131,926 and 133,400 kg, respectively. The results of the other two m