Roster evaluation based on classifiers for the nurse rostering problem

The personnel scheduling problem is a well-known NP-hard combinatorial problem. Due to the complexity of this problem and the size of the real-world instances, it is not possible to use exact methods, and thus heuristics, meta-heuristics, or hyper-heuristics must be employed. The majority of heuristic approaches are based on iterative search, where the quality of intermediate solutions must be calculated. Unfortunately, this is computationally highly expensive because these problems have many constraints and some are very complex. In this study, we propose a machine learning technique as a tool to accelerate the evaluation phase in heuristic approaches. The solution is based on a simple classifier, which is able to determine whether the changed solution (more precisely, the changed part of the solution) is better than the original or not. This decision is made much faster than a standard cost-oriented evaluation process. However, the classification process cannot guarantee 100 % correctness. Therefore, our approach, which is illustrated using a tabu search algorithm in this study, includes a filtering mechanism, where the classifier rejects the majority of the potentially bad solutions and the remaining solutions are then evaluated in a standard manner. We also show how the boosting algorithms can improve the quality of the final solution compared with a simple classifier. We verified our proposed approach and premises, based on standard and real-world benchmark instances, to demonstrate the significant speedup obtained with comparable solution quality.     

Adaptive neighborhood search for nurse rostering

This paper presents an adaptive neighborhood search method (ANS) for solving the nurse rostering problem proposed for the First International Nurse Rostering Competition (INRC-2010). ANS uses jointly two distinct neighborhood moves and adaptively switches among three intensification and diversification search strategies according to the search history. Computational results assessed on the three sets of 60 competition instances show that ANS improves the best known results for 12 instances while matching the best bounds for 39 other instances. An analysis of some key elements of ANS sheds light on the understanding of the behavior of the proposed algorithm.     

New approaches to nurse rostering benchmark instances

This paper presents the results of developing a branch and price algorithm and an ejection chain method for nurse rostering problems. The approach is general enough to be able to apply it to a wide range of benchmark nurse rostering instances. The majority of the instances are real world applications. They have been collected from a variety of sources including industrial collaborators, other researchers and various publications. The results of entering these algorithms in the 2010 International Nurse Rostering Competition are also presented and discussed. In addition, incorporated within both algorithms is a dynamic programming method which we present. The algorithm contains a number of heuristics and other features which make it very effective on the broad rostering model introduced.     

Integer programming techniques for the nurse rostering problem

This work presents integer programming techniques to tackle the problem of the International Nurse Rostering Competition. Starting from a compact and monolithic formulation in which the current generation of solvers performs poorly, improved cut generation strategies and primal heuristics are proposed and evaluated. A large number of computational experiments with these techniques produced the following results: the optimality of the vast majority of instances was proved, the best known solutions were improved by up to 15 % and strong dual bounds were obtained. In the spirit of reproducible science, all code was implemented using the Computational Infrastructure for Operations Research.     

Progress control in iterated local search for nurse rostering

This paper describes an approach in which a local search technique is alternated with a process which ‘jumps’ to another point in the search space. After each ‘jump’ a (time-intensive) local search is used to obtain a new local optimum. The focus of the paper is in monitoring the progress of this technique on a set of real world nurse rostering problems. We propose a model for estimating the quality of this new local optimum. We can then decide whether to end the local search based on the predicted quality. The fact that we avoid searching these bad neighbourhoods enables us to reach better solutions in the same amount of time. We evaluate the approach on five highly constrained problems in nurse rostering. These problems represent complex and challenging real world rostering situations and the approach described here has been developed during a commercial implementation project by ORTEC bv.     

A scatter search methodology for the nurse rostering problem

The benefits of automating the nurse scheduling process in hospitals include reducing the planning workload and associated costs and being able to create higher quality and more flexible schedules. This has become more important recently in order to retain nurses and to attract more people into the profession. Better quality rosters also reduce fatigue and stress due to overwork and poor scheduling and help to maximise the use of leisure time by satisfying more requests. A more contented workforce will lead to higher productivity, increased quality of patient service and a better level of healthcare. This paper presents a scatter search approach for the problem of automatically creating nurse rosters. Scatter search is an evolutionary algorithm, which has been successfully applied across a number of problem domains. To adapt and apply scatter search to nurse rostering, it was necessary to develop novel implementations of some of scatter search's subroutines. The algorithm was then tested on publicly available real-world benchmark instances and compared against previously published approaches. The results show the proposed algorithm is a robust and effective method on a wide variety of real-world instances.     

A comparison of two approaches to nurse rostering problems

Despite decades of research into automated methods for nurse rostering and some academic successes, one may notice that there is no consistency in the knowledge that has been built up over the years and that many healthcare institutions still resort to manual practices. One of the possible reasons for this gap between the nurse rostering theory and practice is that often the academic community focuses on the development of new techniques rather than developing systems for healthcare institutions. In addition, methods suitable for one problem are usually not easily transferable to other problems. In real-world healthcare environments, a personnel manager cannot afford to model a problem and construct a roster using available approaches in order to quantitatively determine which one suits best. There is a lack of criteria for the comparison of approaches to provide a clear picture about their advantages and disadvantages and therefore their suitability to a problem in hand. This paper introduces seven criteria: expressive power, flexibility, algorithmic power, learning capabilities, maintenance, rescheduling capabilities, and parameter tuning, that may offer guidance to researchers and developers of systems for nurse rostering. Two approaches to nurse rostering, which are of very different nature, are evaluated and compared against the introduced criteria. One approach is based on meta-heuristics, while the other employs case-based reasoning.     

Integer programming model extensions for a multi-stage nurse rostering problem

Annals of Operations Research - In the variant of the well studied nurse rostering problem proposed in the Second International Nurse Rostering Competition, multiple stages have to be solved...     

A variable neighborhood search based matheuristic for nurse rostering problems

A practical nurse rostering problem, which arises at a ward of an Italian private hospital, is considered. In this problem, it is required each month to assign shifts to the nursing staff subject to various requirements. A matheuristic approach is introduced, based on a set of neighborhoods iteratively searched by a commercial integer programming solver within a defined global time limit, relying on a starting solution generated by the solver running on the general integer programming formulation of the problem. Generally speaking, a matheuristic algorithm is a heuristic algorithm that uses non trivial optimization and mathematical programming tools to explore the solutions space with the aim of analyzing large scale neighborhoods. Randomly generated instances, based on the considered nurse rostering problem, were solved and solutions computed by the proposed procedure are compared to the solutions achieved by pure solvers within the same time limit. The results show that the proposed solution approach outperforms the solvers in terms of solution quality. The proposed approach has also been tested on the well known Nurse Rostering Competition instances where several new best results were reached.     

A systematic two phase approach for the nurse rostering problem

Nurse rostering is an NP-hard combinatorial problem which makes it extremely difficult to efficiently solve real life problems due to their size and complexity. Usually real problem instances have complicated work rules related to safety and quality of service issues in addition to rules about quality of life of the personnel. For the aforementioned reasons computer supported scheduling and rescheduling for the particular problem is indispensable. The specifications of the problem addressed were defined by the First International Nurse Rostering Competition (INRC2010) sponsored by the leading conference in the Automated Timetabling domain, PATAT-2010. Since the competition imposed quality and time constraint requirements, the problem instances were partitioned into sub-problems of manageable computational size and were then solved sequentially using Integer Mathematical Programming. A two phase strategy was implemented where in the first phase the workload for each nurse and for each day of the week was decided while in the second phase the specific daily shifts were assigned. In addition, local optimization techniques for searching across combinations of nurses’ partial schedules were also applied. This sequence is repeated several times depending on the available computational time. The results of our approach and the submitted software produced excellent solutions for both the known and the hidden problem instances, which in respect gave our team the first position in all tracks of the INRC-2010 competition.     

Local search neighbourhoods for dealing with a novel nurse rostering model

A novel nurse rostering model is developed to represent real world problem instances more accurately. The proposed model is generic in the sense that it allows modelling of essentially different problem instances. Novel local search neighbourhoods are implemented to take advantage of the problem properties represented by the model. These neighbourhoods are used in a variable neighbourhood search and in an adaptive large neighbourhood search algorithm. The performance of the solution method is evaluated empirically on real world data. The proposed model is open to further extensions for covering personnel planning problems in different sectors and countries.     

A hybrid heuristic ordering and variable neighbourhood search for the nurse rostering problem

This paper is concerned with the development of intelligent decision support methodologies for nurse rostering problems in large modern hospital environments. We present an approach which hybridises heuristic ordering with variable neighbourhood search. We show that the search can be extended and the solution quality can be significantly improved by the careful combination and repeated use of heuristic ordering, variable neighbourhood search and back-tracking. The amount of computational time that is allowed plays a significant role and we analyse and discuss this. The algorithms are evaluated against a commercial Genetic Algorithm on commercial data. We demonstrate that this methodology can significantly outperform the commercial algorithm. This paper is one of the few in the scientific nurse rostering literature which deal with commercial data and which compare against a commercially implemented algorithm.     

The nurse rostering problem: A critical appraisal of the problem structure

This paper is concerned with the problem of nurse rostering within hospitals. We analyse a class of four benchmark instances from the nurse rostering literature to provide insight into the nature of the problem. By highlighting the structure of the problem we are able to reduce the relevant solution space. A mixed integer linear programme is then able to find optimal solutions to all four instances of this class of benchmark problems, each within half an hour. Our second contribution is to extend current mathematical approaches to nurse rostering to take better account of the practical considerations. We provide a methodology for handling rostering constraints and preferences arising from the continuity from one scheduling period to the next.     

An estimation of distribution algorithm with intelligent local search for rule-based nurse rostering

This paper proposes a new memetic evolutionary algorithm to achieve explicit learning in rule-based nurse rostering, which involves applying a set of heuristic rules for each nurse's assignment. The main framework of the algorithm is an estimation of distribution algorithm, in which an ant-miner methodology improves the individual solutions produced in each generation. Unlike our previous work (where learning is implicit), the learning in the memetic estimation of distribution algorithm is explicit, that is, we are able to identify building blocks directly. The overall approach learns by building a probabilistic model, that is, an estimation of the probability distribution of individual nurse–rule pairs that are used to construct schedules. The local search processor (ie the ant-miner) reinforces nurse–rule pairs that receive higher rewards. A challenging real-world nurse rostering problem is used as the test problem. Computational results show that the proposed approach outperforms most existing approaches. It is suggested that the learning methodologies suggested in this paper may be applied to other scheduling problems where schedules are built systematically according to specific rules.     

A hybrid model of integer programming and variable neighbourhood search for highly-constrained nurse rostering problems

This paper presents a hybrid multi-objective model that combines integer programming (IP) and variable neighbourhood search (VNS) to deal with highly-constrained nurse rostering problems in modern hospital environments. An IP is first used to solve the subproblem which includes the full set of hard constraints and a subset of soft constrains. A basic VNS then follows as a postprocessing procedure to further improve the IP’s resulting solutions. The satisfaction of the excluded constraints from the preceding IP model is the major focus of the VNS. Very promising results are reported compared with a commercial genetic algorithm and a hybrid VNS approach on real instances arising in a Dutch hospital. The comparison results demonstrate that our hybrid approach combines the advantages of both the IP and the VNS to beat other approaches in solving this type of problems. We also believe that the proposed methodology can be applied to other resource allocation problems with a large number of constraints.     

Modelling and evaluation of time aspects in business processes

There is a need for modelling and performance evaluation techniques and tools for a fast and reliable design of workflow systems. This paper introduces a modelling methodology based on coloured stochastic Petri nets. It allows the integration of control flow, organizational, information related and timing aspects in one modelling framework. The processing delays include stochastic distributions in addition to deterministic times. Several workflows and the effects of constrained shared resources needed for different tasks can easily be described and analysed together. Control flow and organizational aspects are modelled separately in resource and workflow models. These models are automatically compiled into one model, which can then be used for qualitative analysis or performance evaluation. The proposed modelling and evaluation method is supported by the software tool TimeNET. An application example shows its use.     

Crew rostering problem in a public transport company

In this paper, we present an applied study commissioned by Metro Bilbao on how to establish a more egalitarian annual allocation of work to drivers. Task allocation is mixed, with some tasks allocated on a rotating basis and others not. The model proposed is solved as a sequence of four types of integer programming problem. The solution obtained is quasi-optimal: all drivers carry out practically the same tasks over the full year. The main contribution of this paper is its method for combining semi-rotating allocation with a planning time frame divided into five periods of three different types with a workload distributed in a non uniform fashion over the days of the week, and with constraints agreed with employees to obtain an egalitarian solution. This method is being implemented at Metro Bilbao, and Eusko Tren has commissioned a study into a similar method by the authors.     

Towards a reference model for timetabling and rostering

Timetabling and rostering research often starts from particular real world problems. The last two decennia have seen a large number of papers discussing cases, models and approaches. This large body of publications does not presently constitute a structured domain that provides guidelines for addressing particular problem instances, nor does it allow identifying gaps where new research is needed.