Summary The personnel of an organization often has two conflicting goals. Individual employees like to have a good work-life balance, by having personal preferences taken into account, whereas there is also the common goal to work efficiently. By applying techniques and methods from Operations Research, a subfield of applied mathematics, we show that operational efficiency can be achieved while taking personnel preferences into account. In the design of optimization methods, we explicitly consider that these methods should enable the business users to understand and effectively steer the outcomes of these methods. Designing such methods, and applying these to personnel scheduling methods is at the core of the research in this dissertation. The content of this dissertation is summarized in this chapter. Chapter 2: Research Relevance and Outline In Chapter 2, we motivate that employee preferences should be carefully considered in personnel planning and scheduling. In service industries, especially in healthcare, personnel wages are a major part of the operational expenses. Hence, efficient personnel schedules help to control operational expenses. In addition, aging populations imply that on the one hand demand is increasing, and on the other hand that the relative size of the ‘working population’ becomes smaller, which stresses a need for efficient personnel scheduling. In this dissertation, we discuss various operations research methods and practice implementations that address requests and preferences of individual employees on different levels of planning and scheduling. In addition, Chapter 2 gives a short introduction into Operations Research, and provides a brief description of the research environment. Chapter 3: Terminology and Literature Survey Chapter 3 discusses how the literature considers preferences and characteristics of individual employees in personnel planning and scheduling decisions. Furthermore, it introduces a terminology for personnel planning and scheduling decisions, and provides an overview of the various personnel preferences and characteristics that are considered in the literature. Next to this, Chapter 3 outlines how mathematical optimization methods incorporate these preferences and characteristics. Finally, in Chapter 3, we point to some interesting research directions and discuss how the research in this dissertation provides steps in those directions. Chapter 4: Cost-Efficient Staffing under Annualized Hours Chapter 4 studies how flexibility in workforce capacity can be used to efficiently match workforce capacity and demand. Flexibility in workforce capacity is introduced by the annualized hours regime. Annualized hours allow organizations to measure working time per year, instead of per month or per week, thereby allowing organizations to let employees work more hours in one week and less in another. An additional source of flexibility is hiring employees with different contract types, such as full-time, part-time, and min-max, and by hiring subcontractors. In Chapter 4, we propose a mathematical programming formulation that incorporates annualized hours and shows to be very flexible with regard to modeling various contract types. The objective of the model is to minimize salary cost, thereby covering workforce demand, and using annualized hours. The model is able to address various business questions regarding tactical workforce planning problems, e.g., with regard to annualized hours, subcontracting, and vacation planning. In a case study for a Dutch hospital, we demonstrate that applying annualized hours potentially saves up to 5.2% in personnel wages annually. Chapter 5: Staffing under Annualized Hours Using Cross-Entropy Optimization In Chapter 5, a Cross-Entropy optimization implementation is proposed to solve an annualized hours model that is strongly related to the model of Chapter 4. The goal is to select a cost-efficient set of employees that is supposed to cover a given workforce demand, under the annualized hours regime. Our experimental results show that Cross-Entropy optimization is efficient across a broad range of instances, where real-life sized instances are solved in seconds, which significantly outperforms a mathematical programming formulation that is solved with Cplex. In addition, the solution quality of Cross-Entropy closely approaches the optimal solutions obtained by Cplex. Thereby, our Cross-Entropy implementation offers an outstanding method for real-time decision making, for example in response to unexpected staff illnesses, and scenario analysis. Chapter 6: Shift Rostering Using Decomposition: Assign Days Off First Chapter 6 studies a two-phase decomposition approach to solve the shift rostering problem. The first phase creates a days off schedule, indicating working days and days off for each employee. The second phase assigns shifts to the working days in the days off schedule. This decomposition is motivated by the fact that personnel scheduling constraints are often divided in two categories: one that specifies constraints on working days and days off, while the other specifies constraints on shift assignments. To assess the performance of the decomposition approach, we apply it to public benchmark instances, and compare this to an approach that solves the personnel scheduling problem directly. We use mathematical programming to solve the various shift rostering formulations. We also study the extension that includes night shifts, in addition to days off, in the first phase of the decomposition. Chapter 6 presents a detailed results analysis, and analyzes the effect of various instance parameters on the decompositions’ results. In general, the decompositions significantly reduce the computation time and produce good solutions for most instances. Chapter 7: Shift Rostering Using Decomposition: Assign Weekend Shifts First Chapter 7 introduces a shift rostering problem that surprisingly has not been studied in the literature: the weekend shift rostering problem. It is motivated by our experience that employees’ shift preferences predominantly focus on the weekends, since many social activities happen during weekends. The weekend shift rostering problem addresses the rostering of weekend shifts, for which we have designed a problem specific heuristic. In this chapter, we consider the weekend shift rostering problem as the first phase of the shift rostering problem. To complete the schedule, the second phase assigns the weekday shifts using an existing algorithm. We discuss effects of this two-phase approach both on the weekend work schedule and on the schedule as a whole. We demonstrate that our method is effective both on real-life instances and on public benchmark instances. For situations where the weekend work schedule is one of the key determinants of the quality of the complete schedule, our two-phase approach shows to be effective when incorporated in a commercially implemented algorithm. Chapter 8: Shift Rostering from Staffing Levels: a Branch-and-Price Approach In Chapter 8, we outline an approach that creates work schedules directly from staffing levels. This in contrast with many scheduling methods that first create shifts based on staffing levels, and afterwards create work schedules from the set of created shifts. Our proposed approach offers flexibility with respect to incorporating employee preferences in the creation of work schedules. When creating work schedules directly from staffing levels, employee preferences can be considered effictively when defining shifts. To solve the underlying combinatorial optimization model, we compare a Branch-and-Price (B&P) formulation with a mathematical programming formulation. The mathematical programming approach outperforms B&P in most cases, but we believe B&P to be better able to handle extra scheduling and employee preference constraints. Chapter 9: An Iterative Improvement Heuristic to Support Self-Scheduling Chapter 9 studies a self-scheduling application. Self-scheduling is receiving more and more attention in the literature and in practice. With self-scheduling, employees propose their personal work schedule they prefer to work during a given planning horizon. However, these schedules often do not match with the staffing demand as specified by the organization. Chapter 9, presents an approach to support creating feasible work schedules that uses the work schedules proposed by the employees as input and that aims to divide the burden of shift reassignments ‘fair’ among the employees. Computational results are discussed and we indicate how performance indicators of the resulting schedules can be influenced through various model parameters. The presented approach is flexible and easily extendable, since labor rule checks are isolated from the actual algorithm, which makes it easy to include additional labor rules in the approach. Moreover, our approach enables the user to make a trade-off between the quality of the resulting roster and the extent to which the planner is able to track the decisions of the algorithm. Conclusions This dissertation discusses various personnel planning and scheduling applications. Personnel preferences are an important topic throughout the dissertation. Although we reveal several mathematical challenges open for further research, we believe that the most important challenge lies with implementing these methods in practice. We think it is important that the design of these methods enable the business users to understand and effectively steer the outcomes of these methods. The research in this dissertation is based on practical studies or applications, and we encourage additional research in this direction. Many of the algorithm designs explicitly consider that the algorithms should enable business users to understand and effectively steer the outcomes of these methods. We believe that this aspect, combined with the incorporation of employee specific preferences, will be instrumental in the implementation success of personnel planning and scheduling algorithms.
|Award date||22 Nov 2013|
|Place of Publication||Enschede|
|Publication status||Published - 22 Nov 2013|
- Operations research
- Personnel planning
- Personnel scheduling
- Personnel preferences