Staff optimization for time-dependent acute patient flow

The emergency department is a key element of acute patient flow, but due to high demand and an alternating rate of arriving patients, the department is often challenged by insufficient capacity. Proper allocation of resources to match demand is, therefore, a vital task for many emergency departments.Constrained by targets on patient waiting time, we consider the problem of minimizing the total amount of staff-resources allocated to an emergency department. We test a matheuristic approach to this problem, accounting for both patient flow and staff scheduling restrictions. Using a continuous-time Markov chain, patient flow is modeled as a time-dependent queueing network where inhomogeneous behavior is evaluated using the uniformization method. Based on this modeling approach, we recursively evaluate and allocate staff to the system using integer linear programming until the waiting time targets are respected in all queues of the network. By comparing to discrete-event simulations of the associated system, we show that this approach is adequate for both modeling and optimizing the patient flow. In addition, we demonstrate robustness to the service time distribution and the associated system with multiple classes of patients.     

A multiobjective stochastic program for hospital bed planning

In this paper, we study hospital bed capacity management for a set of existing hospitals when the demand for beds is random. We propose a multiobjective stochastic program model to assign beds to hospital departments. We consider three objective functions to be minimized, which are the cost of creation and management of new beds and the number of nurses and physicians working in these hospitals, subject to demand satisfaction of three kinds of health-care specialities. A certainty equivalent program was derived based on a mixture between the chance constrained approach, the recourse approach and the goal programming approach. Empirical results using real data from 157 Tunisian national hospitals are reported.     

Quick response policy with Bayesian information updates

In this paper we investigate the quick response (QR) policy with different Bayesian models. Under QR policy, a retailer can collect market information from the sales of a pre-seasonal product whose demand is closely related to a seasonal product’s demand. This information is then used to update the distribution for the seasonal product’s demand by a Bayesian approach. We study two information update models: one with the revision of an unknown mean, and the other with the revision of both an unknown mean and an unknown variance. The impacts of the information updates under both models are compared and discussed. We also identify the features of the pre-seasonal product which can bring more significant profit improvement. We conclude that an effective QR policy depends on a precise information update model as well as a selection of an appropriate pre-seasonal product as the observation target.     

Robust surgery loading

We consider the robust surgery loading problem for a hospital’s operating theatre department, which concerns assigning surgeries and sufficient planned slack to operating room days. The objective is to maximize capacity utilization and minimize the risk of overtime, and thus cancelled patients. This research was performed in collaboration with the Erasmus MC, a large academic hospital in the Netherlands, which has also provided historical data for the experiments. We propose various constructive heuristics and local search methods that use statistical information on surgery durations to exploit the portfolio effect, and thereby to minimize the required slack. We demonstrate that our approach frees a lot of operating room capacity, which may be used to perform additional surgeries. Furthermore, we show that by combining advanced optimization techniques with extensive historical statistical records on surgery durations can significantly improve the operating room department utilization.     

Clinical and Operational Risk: A Bayesian Approach

In health care organizations (HCOs) adverse events may provoke dangerous consequences on patients, such as death, a longer hospital stay, and morbidity. As a consequence, HCO’s department needs to manage legal issues and economic reimbursements. Governances and physicians are interested in operational (OR) and clinical risk (CR) assessment, mainly for forecasting and managing losses and for a correct decision making. Currently, scientific researches, which are objected to a quantification of CR and OR in HCO, are scarce; absence of regulatory constraints and limited awareness of benefits due to risk management do not provide incentives to elaborate on how risks can be quantified. This paper is aimed at proposing Bayesian methods to manage operational and clinical adverse events in health care. Bayesian Networks (BNs) are useful for assessing risks given end stage renal disease (ESRD) as a context of application; some prior probability distributions are advised for representing knowledge before experimental results and Bayesian utility functions for making the optimal decision. The method is described as from the theoretical as from the empirical point of view, thanks to the health care and haemodialysis department, for this application. The ultimate goal is to introduce a methodology useful for managing operational and clinical risk for haemodialysis patients and departments.      

Bayesian decision-making in inventory modelling

The objective of this paper is to advocate the use of Bayesianmethods in tackling decision problems with limited past data.It is assumed that a Bayesian approach is least likely to besuccessful when there is no information on which to base a meaningfulprior. Here we use a limiting, invariant, form of the conjugateprior distribution to represent this ignorance. The resultsof decisions based on Bayesian methods with this ‘non-informative’prior are compared with those which result from deriving a pointestimate for the unknown parameter. The particular context consideredhere is that of a single-period inventory model with compoundPoisson demand made up of a known demand size distribution butan unknown demand rate. The demand rate is assumed to be highenough for a normal approximation to the compound Poisson distributionto be used, in which case it is possible to analyse the behaviourdirectly. An extension to the multi-period model with zero leadtime is considered briefly. The results lend support to theuse of Bayesian methods, with or without a meaningful prior,for which the analysis and computation are no more complex thanthose required by standard methods.     

Inventory and sales effort management under unobservable lost sales

A retailer needs to make decisions regarding how much to order and how much sales effort to exert in an environment with uncertain demand. One intrinsic complexity in a typical retail environment is caused by the fact that the retailer can obtain information about demand only based on sales, as demand itself is unobservable. Taking a Bayesian approach, Lariviere and Porteus (1999) show that in such a setting a retailer should stock more to increase the probability of an exact demand observation. In this article, we extend their work by allowing the retailer to control both the stocking quantity and sales effort, which can be used to affect demand. We show that their insights with respect to information stalking carry over to this setting. In addition, our model allows gaining a better understanding of optimal sales effort strategies. We find that demand management has a dual role in supporting information gathering: while at the beginning of a product life cycle it is optimal to support learning effects by sharply reducing sales effort, at later stages of the product life cycle an aggressive strategy of increased promotional activities can be used to harvest the information gathered in earlier periods.     

An exact approach for relating recovering surgical patient workload to the master surgical schedule

No other department influences the workload of a hospital more than the Department of Surgery and in particular, the activities in the operating room. These activities are governed by the master surgical schedule (MSS), which states which patient types receive surgery on which day. In this paper, we describe an analytical approach to project the workload for downstream departments based on this MSS. Specifically, the ward occupancy distributions, patient admission/discharge distributions and the distributions for ongoing interventions/treatments are computed. Recovering after surgery requires the support of multiple departments, such as nursing, physiotherapy, rehabilitation and long-term care. With our model, managers from these departments can determine their workload by aggregating tasks associated with recovering surgical patients. The model, which supported the development of a new MSS at the Netherlands Cancer Institute–Antoni van Leeuwenhoek Hospital, provides the foundation for a decision support tool to relate downstream hospital departments to the operating room.     

考虑患者回诊情况下的门诊科室调度优化

目前国内大型三甲医院最突出的问题就是资源稀缺,这些稀缺资源主要包括手术室、大型检查设备以及门诊科室等,造成患者严重的排队现象,从而导致患者等待时间过长,满意度下降。通过对哈尔滨市某大型医院进行调研,收集相关数据,研究在考虑患者回诊(即患者当天做完各项检查后又回到初次检查的门诊科室)情况下,对医生门诊科室的数量进行调度优化。利用排队论中的动态优先级对排队规则进行限定,同时引入前景理论中的价值函数,确立以最小化患者时间感知不满意度为主要目标,最小化医院运营成本为次要目标的多目标优化问题。并且分析了两个目标的权重参数变化对总体满意度的影响。建立相应的数学模型,利用模拟植物生长算法( plant growth simulation algorithm,简称PGSA)进行算法设计,通过MATLAB进行仿真,得出在有限度优先的排队规则下,能够更大程度的降低患者不满意度,同时保证较低的运营成本,证明了此研究的有效性和可行性。     

Modelling high-tech product life cycles with short-term demand information: a case study

Increasing competition and volatile conditions in high-tech markets result in shortening product life cycles with non-cyclic demand patterns. This study illustrates the use of a demand-characterisation approach that models the underlying shape of product demands in these markets. In the approach, a Bayesian-update procedure combines the demand projections obtained from historical data with the short-term demand information provided from demand leading indicators. The goal of the Bayesian procedure is to improve the accuracy and reduce the variation of historical data-based demand projections. This paper discusses the implementation experience of the proposed approach at a semiconductor-manufacturing company; the key test results are presented using product families introduced over the last few years with a comparison to real-world benchmark demand forecasts.     

Bayesian process optimization using failure amplification method

This work is motivated by a problem of optimizing printed circuit board manufacturing using design of experiments. The data are binary, which poses challenges in model fitting and optimization. We use the idea of failure amplification method to increase the information supplied by the data and then use a Bayesian approach for model fitting. The Bayesian approach is implemented using Gaussian process models on a latent variable representation. It is demonstrated that the failure amplification method coupled with a Bayesian approach is highly suitable for optimizing a process with binary data. Copyright © 2010 John Wiley & Sons, Ltd.     

A bi-objective coordination setup problem in a two-stage production system

This paper addresses a problem arising in the coordination between two consecutive departments of a production system, where parts are processed in batches, and each batch is characterized by two distinct attributes. Due to the lack of interstage buffering between the two stages, these departments have to follow the same batch sequence. In the first department, a setup occurs every time the first attribute of a new batch is different from the one of the previous batch. In the downstream department, there is a setup when the second attribute changes in two consecutive batches. The problem consists in finding a batch sequence optimizing the number of setups paid by each department. This case results in a particular bi-objective combinatorial optimization problem. We present a geometrical characterization for the feasible solution set of the problem, and we propose three effective heuristics, as shown by an extensive experimental campaign. The proposed approach can be also used to solve a class of single-objective problems, in which setup costs in the two departments are general increasing functions of the number of setups.     

A COMPOUND POISSON MODEL FOR LEARNING DISCRETE BAYESIAN NETWORKS

We introduce here the concept of Bayesian networks, in compound Poisson model, which provides a graphical modeling framework that encodes the joint probability distribution for a set of random variables within a directed acyclic graph. We suggest an approach proposal which offers a new mixed implicit estimator. We show that the implicit approach applied in compound Poisson model is very attractive for its ability to understand data and does not require any prior information. A comparative study between learned estimates given by implicit and by standard Bayesian approaches is established. Under some conditions and based on minimal squared error calculations, we show that the mixed implicit estimator is better than the standard Bayesian and the maximum likelihood estimators. We illustrate our approach by considering a simulation study in the context of mobile communication networks.     

Capacity allocation for demand of different customer-product-combinations with cancellations, no-shows, and overbooking when there is a sequential delivery of service

We consider a problem where different classes of customers can book different types of services in advance and the service company has to respond immediately to the booking request confirming or rejecting it. Due to the possibility of cancellations before the day of service, or no-shows at the day of service, overbooking the given capacity is a viable decision. The objective of the service company is to maximize profit made of class-type specific revenues, refunds for cancellations or no-shows as well as the cost of overtime. For the calculation of the latter, information of the underlying appointment schedule is required. Throughout the paper we will relate the problem to capacity allocation in radiology services. Drawing upon ideas from revenue management, overbooking, and appointment scheduling we model the problem as a Markov decision process in discrete time which due to proper aggregation can be optimally solved with an iterative stochastic dynamic programming approach. In an experimental study we successfully apply the approach to a real world problem with data from the radiology department of a hospital. Furthermore, we compare the optimal policy to four heuristic policies, of whom one is currently in use. We can show that the optimal policy significantly improves the currently used policy and that a nested booking limit type policy closely approximates the optimal policy and is thus recommended for use in practice.     

Bayesian demand updating in the lost sales newsvendor problem: A two-moment approximation

We consider Bayesian updating of demand in a lost sales newsvendor model with censored observations. In a lost sales environment, where the arrival process is not recorded, the exact demand is not observed if it exceeds the beginning stock level, resulting in censored observations. Adopting a Bayesian approach for updating the demand distribution, we develop expressions for the exact posteriors starting with conjugate priors, for negative binomial, gamma, Poisson and normal distributions. Having shown that non-informative priors result in degenerate predictive densities except for negative binomial demand, we propose an approximation within the conjugate family by matching the first two moments of the posterior distribution. The conjugacy property of the priors also ensure analytical tractability and ease of computation in successive updates. In our numerical study, we show that the posteriors and the predictive demand distributions obtained exactly and with the approximation are very close to each other, and that the approximation works very well from both probabilistic and operational perspectives in a sequential updating setting as well.     

A queuing-based decision support methodology to estimate hospital inpatient bed demand

Hospital inpatient bed capacity might be better described as evolved than planned. At least two challenges lead to this behaviour: different views of patient demand implied by different data sets in a hospital and limited use of scientific methods for capacity estimation. In this paper, we statistically examine four distinct hospital inpatient data sets for internal consistency and potential usefulness for estimating true patient bed demand. We conclude that posterior financial data, billing data, rather than the census data commonly relied upon, yields true hospital bed demand. Subsequently, a capacity planning tool, based upon queuing theory and financial data only, is developed. The delivery mechanism is an Excel spreadsheet. One adjusts input parameters including patient volume and mix and instantaneously monitors the effect on bed needs across multiple levels of care. A case study from a major hospital in Phoenix, Arizona, USA is used throughout to demonstrate the methodologies.     

An integrated queuing and multi-objective bed allocation model with application to a hospital in China

In this paper, a multi-objective decision aiding model is introduced for allocation of beds in a hospital. The model is based on queuing theory and goal programming (GP). Queuing theory is used to obtain some essential characteristics of access to various departments (or specialities) within the hospital. Results from the queuing models are used to construct a multi-objective decision aiding model within a GP framework, taking account of targets and objectives related to customer service and profits from the hospital manager and all department heads. The paper describes an application of the model, dealing with a public hospital in China that had serious problems with loss of potential patients in some departments and a waste of hospital beds in others. The performance of the model and implications for hospital management are presented.     

An efficiency frontier approach for the design of cellular manufacturing systems in a lumpy demand environment

A new multi-objective approach for the cell formation problem in a lumpy demand environment is presented. The objectives addressed in this paper are grouping efficiency and capacity requirements. In lumpy demand the required capacity is affected by demand variability and the correlation between the part types assigned to the cells. We claim that since the required capacity is determined by part types grouping, part type demands variability and their correlation should be taken into consideration as part of the cell formation. This new approach is discussed and formulated as a mixed integer programming model and illustrated by a wide range of typical examples. These examples demonstrate that when using traditional approaches designers do not obtain optimal solutions and may make decisions on the basis of wrong results. The proposed approach helps designers eliminate these problems and produce a reasonable cell design. A genetic algorithm is proposed and examined for designing large-scale systems.     

Strategic information sharing in a supply chain

We consider a two-member supply chain that manufactures and sells newsboy-type products and comprises a downstream retailer and an upstream vendor. In this supply chain, the vendor is responsible for making stock-level decisions and holding the inventory, and the retailer is better informed about market demand. In each period, the retailer receives a signal about market demand before the actual demand is realized, and must decide whether to reveal the information to the vendor, at a cost, before the vendor starts production. We assume that any information that the retailer reveals is truthful. We model the situation as a Bayesian game, and find that, in equilibrium, whether the retailer reveals or withholds the information depends on two things—the cost of revealing the information and the nature of market demand signal that the retailer receives. If the cost of sharing the information is sufficiently large, then the retailer will withhold the information from the vendor regardless of the type of signal that is received. If the cost of sharing the information is small, then the retailer will reveal the information to the vendor if a high demand is signaled, but will withhold it from the vendor if a low demand is signaled. In general, reducing the cost of sharing information and increasing the profit margin of either the retailer or the vendor (or reducing the cost of the vendor or retailer) will facilitate information sharing.     

Confidence-based optimisation for the newsvendor problem under binomial,Poisson and exponential demand

We introduce a novel strategy to address the issue of demand estimation in single-item single-period stochastic inventory optimisation problems. Our strategy analytically combines confidence interval analysis and inventory optimisation. We assume that the decision maker is given a set of past demand samples and we employ confidence interval analysis in order to identify a range of candidate order quantities that, with prescribed confidence probability, includes the real optimal order quantity for the underlying stochastic demand process with unknown stationary parameter(s). In addition, for each candidate order quantity that is identified, our approach produces an upper and a lower bound for the associated cost. We apply this approach to three demand distributions in the exponential family: binomial, Poisson, and exponential. For two of these distributions we also discuss the extension to the case of unobserved lost sales. Numerical examples are presented in which we show how our approach complements existing frequentist—e.g. based on maximum likelihood estimators—or Bayesian strategies.