ABSTRACT
Nowadays, maintenance is more and more considered as an important support function in business with significant investments in physical assets and plays an important role in achieving the organizational goals (Tsang 2002). As a consequence, the measurement of the maintenance performance has become an essential element of strategic thinking of asset owners and managers to be utilized for identifying business processes, areas and departments that need to be improved to achieve the organizational goals (Parida 2006). Over the last two decades, a large number of performance measurement frameworks has been suggested as regards the evaluation of the overall organizational performance. To the contrary, the area of maintenance performance and management is still in need of more future systematic research efforts (Simões et al. 2011). Nowadays, literature on Maintenance Performance Measurement (MPM) models and Maintenance Key Performance Indicators (MKPIs) principally lacks on two aspects. Firstly, proposed MPM frameworks are generic, without considering the business specific environment of the company where these tools should be applied in. Secondly, the available literature mainly proposes a huge list of MKPIs but lacks an agreed-upon methodological approach for selecting specific MKPIs from the listed indicators. With this recognition, the present paper aims at proposing a structured methodology
collected in order to obtain a comprehensive list of MKPIs. Among them, indicators deemed to be the most appropriate to describe the specific maintenance process are selected and then assigned to the hierarchical framework. Then, an Analytic Hierarchy Process (AHP)-based approach (Saaty 1994) is proposed to rank the assigned MKPIs. In particular, the decision maker is asked to supply just pair-wise comparison judgments on which he/she is confident. Namely, incomplete pair-wise comparison matrices could be supplied and then need to be opportunely handled. Finally, a mathematical programming model is formulated and solved in order to select the optimal set of MKPIs by means of which measuring the maintenance performance. Actually, managing a large number of MKPIs can be counterproductive in terms of lack of conciseness of the returned information. In addition, calculating a large number of MKPIs requires a considerable effort in terms of input data to be recorded. As a consequence, MKPIs to be used need to be carefully selected among those assigned so that if on the one hand they synthesize the needed information in the best possible way, on the other hand their number is not too large.
