The uncertainty in infusion durations and non-homogeneous care level needs of patients are the critical factors that lead to difficulties in chemotherapy scheduling. We study the problem of scheduling patient appointments and assigning patients to nurses under uncertainty in infusion durations for a given day. We consider instantaneous nurse workload, represented in terms of total patient acuity levels, and chair availability while scheduling patients. We formulate a two-stage stochastic mixed-integer programming model with the objective of minimizing expected weighted sum of excess patient acuity, waiting time and nurse overtime. We propose a scenario bundling-based decomposition algorithm to find near-optimal schedules. We use data of a major university hospital to generate managerial insights related to the impact of acuity consideration, and number of nurses and chairs on the performance measures. We compare the schedules obtained by the algorithm with the baseline schedules and those found by applying several relevant scheduling heuristics. Finally, we assess the value of stochastic solution.

输液时间的不确定性和患者护理水平的不均匀性是导致化疗调度困难的关键因素。我们研究了在给定日期输液持续时间不确定的情况下安排患者预约和将患者分配给护士的问题。我们考虑了护士的瞬时工作量，以患者的总视力水平和安排患者时的椅子可用性表示。我们制定了一个两阶段随机混合整数规划模型，目标是最小化过度患者敏锐度、等待时间和护士加班时间的预期加权总和。我们提出了一种基于场景捆绑的分解算法来找到接近最优的时间表。我们使用一家大型大学医院的数据来产生与敏锐度考虑影响相关的管理见解，以及护士和椅子的数量对绩效衡量标准。我们将算法获得的时间表与基线时间表以及通过应用几种相关的调度启发法找到的时间表进行比较。最后，我们评估随机解的价值。