Ms Josephine Varney

Office: 666 | Telephone: +61 8 8313 1605

Research seminars

Doctoral thesis

Optimisation in geothermal energy

Honours thesis

Getting the most from a multi-skilled workforce

This study considers the problem of multi-skilling in the Bridgestone Australia Ltd Tyre Factory. The major 'hard' benefit Bridgestone gains from multi-skilling is a decrease in the cost of line downtime that results from absenteeism. Whilst this is a well known and oft-mentioned benefit of multi-skilling, we found no modelling of this in the literature. The modelling in the literature uses multi-skilling to move workers around, to meet situations of changing demand. The situation at Bridgestone is the opposite of this; Bridgestone uses multi-skilling to meet situations of changing supply (i.e. varying levels of insufficient workers) in the face of constant factory demand. Using the Process Area at Bridgestone as our focus plant, it we aimed to find, for a given skill matrix, the minimum cost to the factory from line downtime, resulting from insufficient labour or insufficient skills, due to absenteeism. By strategically varying the input skill matrix we aimed to further understand the main forces driving this cost. As there was no suitable pre-existing model we developed our own model. The classical binary integer linear programming (BILP) problem for Job Allocation forms the basis of our model. Our final model is also a BILP which, subject to a number of constraints re?ecting the structure at Bridgestone, allocates workers to jobs to minimise the cost to the factory due to line downtime. It works across a whole day and includes a constraint which models the between-shift interactions. These result when operators offer to work before and after their shift but, due to overtime restrictions, can only be allocated to one of them. As an input this model requires six groups of workers (representing the six half-shifts in one day at Bridgestone), with each worker having a given skill set. Finally we simulated absenteeism. Each worker's attendance at work was determined by a Bernoulli Trial, with probability equal to the complement of Bridgestone's absenteeism rate. Similarly, each worker's availability to work overtime on the half-shifts where this is possible, was determined by a Bernoulli Trial, with probability 0.5. Implicit in this modelling is the assumption that each worker makes the decision to attend work independently of other workers. For causes such as illness and injury this assumption is quite accurate. The results of this study demonstrate the following driving principles for shift design at Bridgestone: Shifts of varying size result in half-shifts with varying cost levels. This result is independent of the skill level of the shift. This result will help Bridgestone understand some of the line downtime variation it sees between its shifts. Allowing any skill to be totally removed from any one shift, will increase factory cost. When designing its shifts Bridgestone needs to be mindful of the requirement of workers to provide overtime cover on other shifts. If certain jobs cannot be covered by overtime labour, factory cost will increase (assuming the skill mix on other shifts remains unchanged). Multi-skilling is more effective when skills are spread out evenly among workers. Some jobs at Bridgestone require a knowledge of all the jobs below them, creating a workforce with a 'tree-like' structure. This 'tree-like' structure makes an even spread of skills impossible and hence it reduces the benefits of multi-skilling. Wherever possible this structure should be reduced.