This is joint work with Devon K. Barrow and Bahman Rostami-Tabar and is an initial exploration of the benefits of using Multiple Temporal Aggregation, as implemented in MAPA for call centre forecasting. The preliminary results are encouraging. More details in the attached presentation.
With thousands of call centres worldwide employing millions and serving billions of customers as a first point of contact, accurate scheduling and capacity planning of resources is important. Forecasts are required as inputs for such scheduling and planning in the short medium and long-term. Current approaches involve forecasting weekly demand and subsequent disaggregation into half-hourly, hourly and daily time buckets as forecast are required to support multiple decisions and plans. Once the weekly call volume forecasts are prepared, accounting for any seasonal variations, they are broken down into high frequencies using appropriate proportions that mainly capture the intra-week and intra-day seasonality. Although this ensures reconciled forecasts across all levels, and therefore aligned decision making, it is potentially not optimal in terms of forecasting. On the other hand, producing forecasts at the highest available frequency, and aggregating to lower frequencies, may also not be ideal as very long lead-time forecasts may be required. A third option, which is more appropriate from a forecasting standpoint, is to produce forecasts at different levels using appropriate models for each. Although this has the potential to generate good forecasts, in terms of decision making the forecasts are not aligned, which may cause organisational problems. Recently, Kourentzes et al. (2014) proposed the Multiple Aggregation Prediction Algorithm (MAPA), where forecasting with multiple temporal aggregation (MTA) levels allows both accurate and reconciled forecasts. The main idea of MTA is to model a series at multiple aggregation levels separately, taking advantage of the information that is highlighted at each level, and subsequently combine the forecasts by using the implied temporal hierarchical structure. Athanasopoulos et al. (2017) proposed a more general MTA framework than MAPA, defining appropriate temporal hierarchies and reconciliation mechanisms, and thus providing a MTA forecasting framework that is very flexible and model independent, while retaining all the benefits of MAPA. Given the high frequency, multi-temporal nature of the forecast requirements and the subsequent planning associated with call centre arrival forecasting, MTA becomes a natural, but yet unexplored candidate for call centre forecasting. This work evaluates whether there are any benefits from temporal aggregation both at the level of decision making as well as at the level of aggregation in terms of forecast accuracy and operational efficiency. In doing so, various methods of disaggregation are considered when the decision level and the forecasting level differ, including methods which results in reconciled and unreconciled forecasts. The findings of this study will contribute to call centre management practice by proposing best approaches for forecasting call centre data at the various decision levels taking into account accuracy and operational efficiency, but will also contribute to research on the use of temporal hierarchies in the area of high frequency time series data.