University of Twente Student Theses

Login

Let the weekend begin!:a solution for solving the weekend scheduling problem for ORTEC Harmony

Versteegh, Frédérique (2009) Let the weekend begin!:a solution for solving the weekend scheduling problem for ORTEC Harmony.

[img] PDF
788kB
Abstract:Introduction: This research is performed within the Product Knowledge Center of ORTEC. ORTEC is one of the largest providers of advanced planning and optimization software solutions and consulting services. We study the assignment of weekend shifts in ORTEC's decision support software solution for workforce scheduling, called ORTEC Harmony. The assignment of weekend shifts is of great importance to customers, as weekends have an impact on the social life of employees and thus on employee satisfaction. In addition, unfulfilled demand in weekends is hard to cover and expensive, as irregularity allowances have to be paid for those hours. Harmony has the functionality to automatically create schedules. In a case study we analyze three real-life cases from practice that use this functionality. These customers use a two-step approach to schedule: weekend shifts as a priori step, followed by the remaining shifts. The current alternatives in Harmony to create satisfying schedules are very time-consuming. Even the best approach (varies per user) regularly leads to unfulfilled weekend demand and does not always satisfy user requirements. Users prefer an equitable schedule in which every employee works approximately the same number of weekend shifts and shift types. - Research objective: The objective of this research is to improve the assignment of weekend shifts such that unfulfilled demand is minimized and user requirements are satisfied as much as possible. The problem addressed by this objective is referred to as the Weekend Scheduling Problem. - Solution approach: We model the Weekend Scheduling Problem as a Quadratic Integer Programming model. For this problem we propose a stand-alone solution, the Weekend Planner. The Weekend Planner can be embedded in Harmony, both as an individual planner and as an element of the current optimization engine. The Weekend Planner assigns weekend shifts to employees using an intelligent form of list scheduling. Working a shift means working a specific shift type in a specific weekend, for the whole weekend. For the selection of shifts and resources, the Weekend Planner uses selection rules based on flexibility and busyness. First, the least flexible shift is selected. In case of a tie, the shift is selected based on randomness. Second, the least busy and least flexible resource is selected. Again, in case of a tie, the resource is selected based on randomness. Constraints on availability and on the maximum allowed number of weekends to work are taken into account. The creation of one schedule ends with a local search method. Random sampling is used to create diversity in the resulting schedules. - Results: The Weekend Planner is tested on three real-life cases from practice and on 400 random instances. Randomly generated instances with parameter ranges typically seen in practice are used to test the robustness and sensitivity of the Weekend Planner. Computational experiments show that the Weekend Planner solves 89% of the instances - i.e., all shifts are assigned - and solves 67.5% of the instances to optimality. The results deviate on average 3.8% from optimal. The Weekend Planner is most sensitive for cyclical schedules and schedules in which every employee needs to work its maximum number of weekends, in other words a tight schedule. Computational experiments on the three real-life cases show that the Weekend Planner outperforms Harmony on all aspects and improves schedules by 15% to 400%. - Conclusions: The Weekend Planner is a suitable method for solving the Weekend Scheduling Problem. The Weekend Planner decreases the unfulfilled demand and delivers more equitable schedules than Harmony does. The Weekend Planner focuses on weekend related constraints and therefore is a user-friendly and not time-consuming method. - Recommendations: We recommend to use a more extensive local search method in the Weekend Planner, which is already available in Harmony, to decrease the deviation from optimal and decrease the number of unassigned shifts, if any. Finally, we recommend to implement the Weekend Planner in Harmony.
Item Type:Essay (Master)
Clients:
ORTEC
Faculty:BMS: Behavioural, Management and Social Sciences
Subject:85 business administration, organizational science
Programme:Industrial Engineering and Management MSc (60029)
Link to this item:http://purl.utwente.nl/essays/60656
Export this item as:BibTeX
EndNote
HTML Citation
Reference Manager

 

Repository Staff Only: item control page