Yannick Kergosien

A tabu search heuristic for the dynamic transportation of patients between care units

The problem studied in this paper stems from a real application to the transportation of patients in the Hospital Complex of Tours (France). The ambulance central station of the Hospital Complex has to plan the transportation demands between care units which require a vehicle. Some demands are known in advance and the others arise dynamically. Each demand requires a specific type of vehicle and a vehicle can transport only one person at a time. The demands can be subcontracted to a private company which implies high cost. Moreover, transportations are subject to particular constraints, among them priority of urgent demands, disinfection of a vehicle after the transportation of a patient with contagious disease and respect of the type of vehicle needed. These characteristics involve a distinction between the vehicles and the crews during the modeling phase. We propose a modeling for solving this difficult problem and a tabu search algorithm inspired by Gendreau et al. (1999). This method supports an adaptive memory and a tabu search procedure. Computational experiments on a real-life instance and on randomly generated instances show that the method can provide high-quality solutions for this dynamic problem with a short computation time.

Metaheuristic algorithms for solving two interconnected vehicle routing problems in a hospital complex

This paper addresses a real-life logistic problem arising in the hospital complex of Tours (France). The two-level vehicle routing problem has time windows, a heterogeneous fleet, and multi-depot, multi-commodity and split deliveries. The first level concerns the routing problem for a fleet of vehicles serving several hospital units that delivers medicines, clean linen, meals, various supplies, patient files and picks up waste and dirty linen. The second level concerns the problem of routing employees between buildings within a large hospital unit. Both levels are interconnected. In addition, decisions about sizing and planning the teams of drivers and warehouse employees have to be made. Two metaheuristic algorithms are proposed to solve the entire problem: a genetic algorithm and a tabu search. The algorithms are tested on 100 instances, randomly generated on the basis of real-life instances.

A generic and flexible simulation-based analysis tool for EMS management

Emergency medical services (EMS) are dedicated to provide urgent medical care to any person requiring it and to ensure their transport to a hospital or care facility, if required. Moreover, in many contexts, EMS also have to provide transportation services for patients need to go from one hospital to another or between their home and the hospital. For such organisations, efficient strategies for managing the ambulance fleet at their disposal have to be selected, but the highly random and dynamic nature of the system under study makes this a challenging task. Most of the published studies which have considered these issues have done it focusing on a specific EMS context, one city or one territory for instance. However, it is possible to identify several common characteristics and processes from one EMS context to another. This is the purpose of the generic discrete event simulation-based analysis tool proposed here, which can be adapted to a wide range of EMS contexts. In particular, it explicitly considers the two types of tasks that can compose the mission of an EMS: serving emergency requests and providing transports between care units/hospitals/patients’ homes.

An empirical comparison of relocation strategies in real-time ambulance fleet management

We define four fleet management strategies and model them using a common framework.We report extensive simulation experiments to analyze the proposed strategies.Our empirical study confirms that dynamic strategies dominate static ones.Improvements are obtained at the expense of significant relocations costs. In order to ensure an adequate service to the population, Emergency Medical Services (EMS) rely on a given number of ambulances strategically located over the territory they serve. The arrival of calls to EMS being highly uncertain and dynamic, it may happen that at some point, the vehicles available to respond to these calls no longer cover properly all regions, even if the coverage was carefully planned initially. Relocation of ambulances may therefore be required during the day in order to achieve better performances. Some models tackling relocation have been proposed in the literature and it has been shown that using such strategies can help to improve overall performances. However, relocation generates movements that produce undesirable consequences from both economical and human resources management standpoints. Questions therefore arise: Is the relocation worth the effort? And if so, what form should it take? Unfortunately, this issue has not been investigated much up to now. This study thus focuses on evaluating and analyzing relocation strategies, and reports extensive simulation experiments allowing to analyze the performance of these strategies when the system faces different levels of workload. Our empirical study confirms that dynamic strategies dominate static ones and quantifies the improvements achieved with respect to service level, but also shows that such improvements are obtained at the expense of significant relocation costs.

A Benders decomposition-based heuristic for a production and outbound distribution scheduling problem with strict delivery constraints

The problem addressed in this paper is from a chemotherapy production and delivery environment, where production and delivery are strongly connected problems. Independent jobs have to be prepared by pharmacy technicians working in parallel. These jobs represent pouches of injectable chemotherapy preparations. The production process corresponds to a classic parallel machine scheduling problem. Then, the jobs must be delivered to the patients by a given due date. Only one person ensures all the deliveries, making several trips between the pharmacy production unit and the patient locations. We model this step as a multi-trip traveling salesman problem, where only one salesman can make more than one trip. The objective to minimize is the maximum tardiness of delivery. In addition to the constraints that link the two problems, some constraints related to the chemical stability of chemotherapy drugs have to be taken into account: The time between the production starting time and the date the treatment is administered to the patient (here, the delivery time) cannot exceed the stability duration, as the drug may otherwise become dangerous or ineffective for the patient. Due to these constraints, the problem is more difficult to solve. The proposed resolution method in this paper is a Benders decomposition-based heuristic that makes it possible to find feasible solutions and lower bounds. The advantage of the Benders decomposition approach is that this method exploits the structure of the problem, which can be easily decomposed into two stages. Computational experiments are conducted, and a comparison with a direct exact resolution shows the efficiency of this approach.

A Routing Problem for Medical Test Sample Collection in Home Health Care Services

Health care organizations are increasingly turning towards home care solutions to provide services to the population under their jurisdiction. Among these services, medical test sample collection (in particular blood and urine) is a highly demanded service by medical doctors. Providing this type of service clearly requires efficient coordination and planning of appointments for patients and healthcare personnel. This management involves the solution of a type of vehicle routing problem that is very complex given the large number and particular nature of the constraints that need to be considered: personnel schedules, patients preferences, maximum transport delay for some blood samples, etc. In this paper, we propose an integer linear programming formulation and different metaheuristics to solve this special vehicle routing problem. Experimental results on randomized instances were performed in order to select the best method to be integrated into a decision support tool and test it in on real data. This study is the result of collaboration with the blood collection service of the Center for Health and Social Service of Laval in the province of Quebec (CSSS-Laval).

Managing a Fleet of Ambulances to Respond to Emergency and Transfer Patient Transportation Demands

Organizations of pre-hospital emergency medical services have as first mission to provide medical assistance to patients including transport to medical centers if necessary, and a second one that concerns the transfers of patients from one medical facility to another one. Most organizations in Canada and in North-America use two independent fleets to perform these missions. Although operating two separated fleets seems easier to do, it appears to be less efficient than an integrated fleet management approach able to deal with both types of demands. Taking this into consideration, this paper aims to design and evaluate the performance of three management approaches. This is a very challenging problem, since it involves solving simultaneously two difficult vehicle routing problems: an ambulance relocation problem and a dial a ride problem.