论文部分内容阅读
针对妇产科孕妇产检多重入、周期长的特点,建立以最小化违背孕妇偏好的惩罚成本和医生的加班成本为目标的孕妇产检时间指派优化随机规划模型。利用蒙特卡洛仿真的方法模拟了多个场景下未来孕妇到达的不确定性及其偏好的不确定性,将随机规划模型转化成了线性规划模型。为了减少计算时间,基于短视策略提出了只考虑当周到达孕妇的线性规划模型和贪婪算法。数值实验表明,基于多场景的随机规模模型得到的结果最好(总成本最低),但计算时间长;而只考虑当周到达的线性规划模型方法和贪婪算法计算时间较短,但求解精度稍差。参数的敏感度分析发现,到达率越高、孕妇偏好分布越集中,总成本就会越高。
In view of the characteristics of maternity multiple check-in and long cycle in maternity and gynecology, a stochastic programming model of maternity time assignment was established to minimize the penalty cost of violating the preferences of pregnant women and overtime costs of doctors. The Monte Carlo simulation method was used to simulate the uncertainties of future arrival of pregnant women in multiple scenarios and their preferences. The stochastic programming model was transformed into a linear programming model. In order to reduce the computational time, a linear programming model and a greedy algorithm are proposed based on the short-sighted strategy to consider only pregnant women arriving in the week. Numerical experiments show that the stochastic scale model based on multi-scenario has the best result (the lowest total cost), but it takes a long time to compute. However, only the linear programming model and the greedy algorithm which arrive in the week are short, difference. Sensitivity analysis of the parameters found that the higher the arrival rate, the more concentrated the preference of pregnant women, the total cost will be higher.