Optimisation of resources deployment in a call centre by using stochastic data in simulation models

In recent years, call centres have been considered as an integral part of the modern businesses, since they play an important role in providing service delivery functions to their customers. A well-managed call centre, therefore, is crucial to ensure high level of customer satisfaction in today's competitive market. In order to achieve a high standard, managers of call centres face a very difficult set of challenges. At the top level, they must strike a balance between two powerful competing interests: low operating costs and high service quality. On a day-to-day basis, while simultaneously keeping low costs and high service quality, those managers must also employ appropriate techniques and tools in order to evaluate the true performance of their operations accurately. Such tools play a vital role in understanding the current system performance, evaluation of any proposed enhancement scenarios, and optimising operations management decisions under any unexpected operating conditions. One of traditional operations management challenges for call centre managers is to tackle the multi-period human resources allocation problem. In this thesis, the staffing and staff scheduling decisions in single-skill inbound call centres were studied. These decisions are normally made under strict service level constrain in the presence of highly uncertain operations and demand of call centre services. Neglecting such uncertainty may lead to unrealistic decisions. The objective of this research thesis was to propose a framework to enhance the call centre performance through taking realistic optimal staffing and scheduling decisions. Realistic optimisation requires realistic modeling (evaluation) of call centre operations which is the main focus and contribution of this research. The proposed framework has combined statistical, simulation, and Integer Programming (IP) techniques in achieving realistic optimisation. The framework begins by developing stochastic statistical data models for call centre operations parameters which are divided into service demand (arrival volumes) and service quality (service times, abandonment volumes, and patience time) parameters. These data models are then fed into a simulation model which was developed to determine the minimum staffing levels in daily an-hour periods. Finally, these staffing levels are considered as input to an IP model that optimally allocates the service agents to the different operating shifts of a typical working day. Application of the proposed framework to a call centre in Libya will also be presented to illustrate how its staffing and scheduling decisions could be improved by using the model.