Abstract: In many health examination agencies,the unreasonable scheduling mechanisms make customers stay in the queue for longer time.This paper presents the study on the scheduling of health examination customers under stochastic service time.Using an appointment scheduling strategy,the order of examination items for each customer is optimized,and on this basis,the paper proposes a two-stage stochastic simulation optimization algorithm that includes rough simulation evaluation and accurate simulation evaluation.The algorithm employs the idea of ordinal optimization,and takes the multi-population genetic algorithm based on affinity evaluation as an iterative optimization strategy.In addition,the improved Optimal Computing Budget Allocation(OCBA) method is used to eliminate the influence of the super individuals,forming a global and adaptive optimal allocation mechanism of simulation resources.The experimental results show that the proposed algorithm obtains a better scheduling solution than the discrete events without scheduling and the events using heuristic scheduling rules to determine the examination order.

Key words: Appointment scheduling, multi-population genetic algorithm, simulation optimization algorithm, Ordinal Optimization(OO), Optimal Computing Budget Allocation(OCBA)

摘要： 针对体检机构顾客排队等待时间长的问题，研究随机服务时间下的体检顾客调度，采用多人时间槽预约策略，并在预约调度策略的基础上优化每位顾客的体检项目顺序，提出一种包含粗糙仿真评估和精确仿真评估两阶段随机仿真优化算法。运用序优化思想将基于亲和度评估的多种群遗传算法作为迭代优化策略，并利用改进的最优计算量分配方法排除超级个体的影响，形成仿真资源的全局和自适应优化分配机制。实验结果表明，与不进行任何调度及使用体检顺序启发式调度规则的离散事件结果相比，该算法获得了更好的调度解。

关键词: 预约调度, 多种群遗传算法, 仿真优化算法, 序优化, 最优计算量分配

CLC Number: 

• TP391