Abstract
In this paper, a Positron Emission Tomography/Computed Tomography (PET/CT) examination scheduling problem considering multi-stage processes is studied. Before the actual examination process, imaging agents (a drug with radioactivity) need to be injected into patients. The radioactivity of the imaging agents continuously decays, which results in the required dose by patients increasing with time, i.e., the later the injection time, the more imaging agents need to be prepared for the patients at the beginning. Considering the imaging agents are expensive and non-storable, the studied problem is to determine the start time of the examination service and injection time for the patients, to minimize the total dose of purchased imaging agents. An integer programming model and a set partitioning model are formulated for this problem. A variable neighborhood search heuristic is proposed, in which a scheduling rule based on some derived optimal properties is embedded as one of the search operators. Computational experiments show that the proposed algorithm can obtain near-optimal solutions in a short time, and moreover find much better results than the commonly used First Come First Service (FCFS) rule in most medical institutions, i.e., our approach’s results need much fewer required dose of the imaging agents, and hence can save a lot of costs for the medical institutions.
View more >>