Batterley Algorythm

The Batterley Algorithm is a workforce optimisation algorithm used for scheduling, especially within call centres. The Batterley algorithm transforms a workforce requirement graph into and optimised and discrete set of ‘shifts’ which may be further refined to return an optimal set of allocated shifts within a finite timeline.

The Batterley Algorithm is based on a set of data manipulation steps that may be performed manually, or implemented as a computer algorithm.

Overview

The Batterley Algorithm extends Erlang by providing further optimization after the calculation of required/forecast agent levels using tools such as those defined by Erlang C.

i.e. While Erlang may be used to calculate the number of agents required to staff a call centre within discrete time periods, Erlang does not provide a solution for ‘filling’ the forecasted timeslots, for those discrete time periods, with staff. The Batterley Algorithm takes as input the output of forecasting calculations (such as Erlang C) and optimally defines work ‘shifts’ (as output) that match the input workforce requirement curve. Further refinements of the shifts are then made to implement the preferred and regulatory workforce allocation requirements that operate within the scheduling organisation.

Optimisation using the Batterley Algorythm

The primary steps involved in using the Batterley Algorithm for workforce optimisation within call centres may be summarised as follows:

1. Generate workforce requirements based on forecast data and Erlang;

2. Use the output of Erlang as input to Batterley (to define initial ‘shifts’);

3. Apply the preferred and regulatory workforce allocation rules to the output of Batterley to refine ‘shift’ lengths; and

4. Allocate staff to the defined shifts within the resultant schedule;

Steps within the Batterley Algorithm

The following is a generalised version of the Batterley Algorithm, which excludes exceptions caused by ‘plateaus’ which are defined within the formalisation of the algorithm.

1. Take the initial graph as output by workforce forecasting methods (such as Erlang C) and identify the peeks and troughs within the graph. In general, you operate on a discrete time range within this step of 1 day; more specifically, ‘greater than the length of a shift’ and ‘no more than 1 day’;

2. Identify the first ‘peak’ or ‘trough’, whichever comes first;

3. Identify the next peak or trough, and ‘invert’ that section of the graph so that when added to the previous section of the graph, the length of the shifts is approximated as the ‘average’ of the shifts that may logically fit within the combined sections of the graph.

4. Repeat step 3) while logically substituting the section of the graph identified within step 3) for the section as identified within step 2) (i.e. use the same method along the rest of the workforce requirement graph).

5. When you have reached the end of the graph, the result is the most optimal set of shifts that can be defined for the forecast data, but that which may not be suitable for the preferred and regulatory workforce allocation requirements applicable to the scheduling organisation. From this point on, the shifts may be manipulated past the optimisation of the Batterley Algorithm.

Benefits

The following are the benefits provided by the Batterley Algorithm:

1. Because the Batterley Algorithm only requires a single pass of the output of forecast workforce requirements to provide the most optimal fit those requirements, an optimised staff schedule can be drafted within a reasonable timeframe.

2. Forecasting methods, such as Erlang C, provide no further insight into how to optimise workforce allocations to workforce requirements for discrete time frames. The Batterley Algorithm provides ‘the’ most optimal set of defined shifts for the forecast data, from which further optimisation is trivialised by being able to extend, shorten or delete the required shifts defined by using the Batterley Algorithm.

3. Adding of required shifts after application of the Batterley Algorithm is rarely, if ever, required. This because the most optimal fit for the forecast workforce requirements has already been performed.

Background

The Batterley Algorithm is attributed to Sara Batterley of Sydney, Australia, is now implemented within workforce optimisation software.

Significantly, Batterley was able to perform the calculations required to turn ‘workload forecast data’ into ‘defined shifts’ for a call centre of some 480 staff while only using pencil and paper. In formulating the method, mental calculations were performed mentally by Batterley, and it was not until working with a software company in the late 1990s that a formal definition and implementation of the algorithm was specified and adopted.

See Also

Erlang

Main Menu