Learning and forgetting-based worker selection for tasks of varying complexity

Abstract

This paper presents an approach for selecting workers for tasks of varying complexity based on individual learning and forgetting characteristics in order to improve system productivity. The performance of a learning and forgetting-based selection (LFBS) policy is examined using simulation and compared to a baseline policy representing criteria used in practice. The effects of factors including worker redundancy and task-tenure on productivity are also examined in the environment of continuously staffed independent tasks. Results demonstrate that the LFBs policy significantly improves productivity relative to common practice and suggests that lower levels of redundancy and shorter tasktenures tend to mitigate some of the negative effects of forgetting.

Selection policies for a multifunctional workforce

“…With MaxiMax O and MaxiMin O, workers are ranked based on knowledge of all four L/F parameters, and estimates the expected output from each task. With MaxiMax P and MaxiMin P, we rely solely on the single prior expertise parameter, p. We first note that, similar to results from Nembhard and Osothsilp (2005), these partially informed policies perform no worse than the baseline, and in some cases significantly outperform the baseline. Secondly, while Nembhard and Osothsilp suggest that basing worker selection and assignment on L/F characteristics can provide the opportunity for improvement, we further demonstrate that several straightforward policies, which do not require instance specific optimisation, can also provide such opportunities.…”

“…We begin with a pool of n workers based on empirical data from the literature (Shafer, Nembhard, and Uzumeri 2001;Nembhard and Osothsilp 2005). The data span a 6-month period from the final inspection stations of a car radio assembly process.…”

“…Norman et al (2002) developed a team-based model for worker assignment in manufacturing cells with both technical and human skills, demonstrating that when skills are explicitly taken into account in both training plans and assignment strategies, the performance of the cell will improve significantly. Nembhard (2001) proposed a heuristic approach for selecting workers for tasks based on individual learning rates, and Nembhard and Osothsilp (2005) presented an approach for selecting workers for tasks of varying complexity based on individual L/F characteristics in order to improve the system productivity. Digiesi et al (2009) showed that dynamic worker behaviour influences the queuing behaviour of flow lines with a model that accounts for fatigue and learning effects.…”

The Effect of Worker Heterogeneity in Learning and Forgetting on System Productivity

“…As suggested in (2), the exponential L/F model has one weakness that it does count the cumulative number of periods for learning and forget- [7] or the hyperbolic recency L/F model [13,18,16].…”

Multi-period dynamic technician routing and scheduling problems with experience-based service times and stochastic customers

“…Reviews of the literature on learning curves can be found in Dar-El (2000), Jaber and Sikström (2004), Jaber (2006), and Anzanello and Fogliatto (2011). Because of the availability of distributions from which to generate workforces, in this research, we use the hyperbolic learning model described in Nembhard and Osothsilp (2005). We note that, while we employ the hyperbolic learning model, most learning curves have simi-lar shapes and would support conclusions similar to those discussed in Section 4.4.…”