Abstract
This paper provides a simulation model for scheduling service task operations and distributing related human resources in dispersed work centres. The managerial concern for the minimisation of temporal overhead costs of task operations in the face of fluctuating, short-term service demands is examined under restrictions imposed by resource availability, work hour flexibility and task-backlog fulfilment. Scheduling strategies are developed directly from the constrained reduction of temporal overheads of appointment and release operations in distributed, non-interlinked work centres. To ensure the model’s structural validity, simulated task backlogs are adjusted to the actual backlog-reducing procedures in real applications. The model provides means for setting up balanced work schedules that can greatly lower temporal overheads of appointment and release operations if workers are selected in accordance with compatible time availability and task qualifications. Direct comparisons of worker productivities in the different centres can also be made, allowing managers to locate bottleneck points of service operations when productivity falls short of desired expectations. The robustness of the model is ensured by finding significant parameter domains through Monte Carlo simulations, centred on data points collected from real-time demand functions in actual service operations.
Notes
1. In fact, the present model emerged out of a few applications of this sort in which the author of this paper had the opportunity to serve as a production modelling and implementation advisor a few years ago.
2. Although the majority of the employees at the work centres are engineering graduates equipped with the relevant technical vocabulary, they speak English only as a second language, which is heavily loaded with culture-specific idioms and expressions.
3. Onsite managers of CompanyX who work in US offices only.
4. The flow time in the model represents a parametric variable having no physical significance. By contrast, the actual periodic interval of work, denoted by (Δt), is a physically measurable quantity. Hence, the start and end points of the time domain can be set to any suitable values appropriate to a specific problem.
5. This was one of the initial applications of the model mentioned previously (designated here by the generic name ServSol). The other applications that employed the model were based on the same basic principles, although the implementation details varied somewhat.
6. Recall that these are computed as the average of the relative percent difference between actual data point and the corresponding simulation value.
7. These tests are performed for τ2 only.