In radiotherapy, treatment needs to be delivered in time. Long waiting times can result in patient anxiety and growth of tumors. They are often caused by inefficient use of radiotherapy equipment, the linear accelerators (LINACs). However, making an efficient schedule is very challenging, especially when we have multiple types of patients, having different service requirements and waiting time constraints. Moreover, in radiotherapy a patient needs to go through a LINAC multiple times over multiple days, to complete the treatment. In this paper we model the radiotherapy treatment process as a queueing system with multiple queues, and we propose a new class of scheduling policies that are simple, flexible and fair to patients. Numerical experiments show that our new policy outperforms the commonly used policies. We also extend the policy to an adaptive one to deal with unknown and fluctuating arrival rates. Our adaptive policy turns out to be quite efficient in absorbing the effects caused by these changes. Due to the complexity of our problem, we select the parameters of the policies through simulation-based optimization heuristics. Our work may also have important implications for managers in other service systems such as call centers.

中文翻译：

---

放疗计划的自适应优先级策略

在放射治疗中，需要及时进行治疗。漫长的等待时间可能导致患者焦虑和肿瘤生长。它们通常是由于放疗设备线性加速器（LINAC）使用效率低下引起的。但是，制定有效的计划非常具有挑战性，特别是当我们有多种类型的患者，具有不同的服务要求和等待时间限制时。此外，在放射疗法中，患者需要在数天内多次经历LINAC，才能完成治疗。在本文中，我们将放射治疗过程建模为具有多个队列的排队系统，并提出了一种简单，灵活且对患者公平的新型调度策略。数值实验表明，我们的新策略优于常用策略。我们还将政策扩展为一种适应性政策，以应对未知且波动的到达率。事实证明，我们的适应性政策非常有效地吸收了这些变化所造成的影响。由于问题的复杂性，我们通过基于仿真的优化启发法来选择策略的参数。我们的工作也可能对其他服务系统（例如呼叫中心）的经理产生重要影响。