Multicriteria model for maintenance benchmarking

• This article describes the first multicriteria model to apply competitive and generic maintenance benchmarking.
• Discrete probability distributions are used so this provides a significant improvement over the use of industrial averages.
• The multicriteria model uses fuzzy AHP to include the ambiguities and uncertainties involved in the judgment.
• Discrete probability distributions were created in different industrial sectors.
• The final utility of an alternative is calculated by a multi-measure utility function.

Maintenance offers great scope for improvement in organizations because historically it has not been considered to have a significant effect on competitiveness. The use of benchmarking therefore would allow companies to identify aspects of maintenance functions which are susceptible of improvement when compared with other organizations. This article describes a multicriteria model constructed by a fuzzy analytic hierarchy process (AHP), utility functions and discrete probability distributions for the application of competitive and generic benchmarking in small and medium enterprises. The multicriteria model uses fuzzy AHP to assign weightings to the decision criteria and subcriteria in such a way as to include the ambiguities and uncertainties involved in the judgments of the decision makers. These weightings are converted into utility functions which will then allow the final utility of an alternative to be calculated via a multi-measure utility function. Using data from a survey carried out by the Spanish Maintenance Association, a series of discrete probability distributions were created from the appearance probability at each level of a descriptor in distinct industrial sectors. A Monte Carlo simulation was used to obtain a global maintenance assessment for each industrial sector and company, incorporating uncertainty into the results. The model described provides a global comparison of the state of maintenance in a company according to these criteria with respect to its sector and other sectors with better results. The model has been tested with a pharmaceutical company, a manufacturer of building materials and a sports center.